Although Duolingo isn’t a translation app, Apple’s latest innovation poses an existential threat to the edtech app’s business model. With 47.7 million daily active users (“DAU”) and 46% y/y subscription revenue growth during Q2 2025, Duolingo relies heavily on necessity-driven learnings looking to master new languages for travel and work settings. Yet Apple’s free, instant live translation solution could render such necessity-driven learning obsolete, jeopardizing Duolingo’s subscriptions and DAU growth. Apple’s live translation innovation also threatens iOS user demand for Duolingo lessons – a user cohort that currently accounts for 71% of Duolingo’s revenues.

As AI-driven competition intensifies, the Duolingo stock faces risks of a further valuation reset despite it already being 50% lower from its May peak.

AirPods Pro 3 Live Translation: Beyond Traditional Tools

Unlike text-based apps like Google Translate or Apple Translate, the AirPods Pro 3’s Live Translation feature redefines language barriers in conversation. Users can speak into their AirPods in their native language, and the translated audio will play through the listener’s earpiece, with text transcripts also displayable on an iPhone.

Adoption for the AirPods Pro 3 is expected to be strong. The AirPods currently represent Apple’s best-selling accessory with a 47% penetration rate on the iOS installed base. A strong upgrade cycle is also expected, since the AirPods Pro 3 represents the line-up’s first refresh in two years. The scale of the AirPods’ reach threatens Duolingo as well, as 71% of its revenues currently come from iOS users (15.5% from Google Android users).

Duolingo Subscriptions and DAUs Under Siege

The conversational focus of Apple’s newest live translation feature undermines Duolingo’s core value – learning languages for practical needs. Necessity-driven learners currently fuel Duolingo’s 10.9 million paid subscribers and 46% y/y subscription revenue growth in Q2, given their commitment to daily lessons for fluency. By offering instant translation, the AirPods Pro 3’s live translation feature could erode this cohort’s motivation for committed learning, leaving Duolingo reliant on less sticky leisure learners. Yet these casual users, which represent the bulk of Duolingo’s 47.7 million DAUs, drive slower growing ad and in-app purchase revenues instead of the higher-margin subscription sales critical to Duolingo’s earnings outlook.

The Duolingo stock’s sensitivity to AI-driven competition is reflected in the 50% decline from its May 2025 peak. This is compounded by investors’ growing concerns about Duolingo’s decelerating DAU growth:

• Q3 2024: 54% y/y growth

• Q4 2024: 41%

• Q1 2025: 49%

• Q2 2025: 40% (vs. Q2 2024: 59% and Q2 2023: 62%)

Duolingo emphasizes DAUs over monthly users, noting “language learning is a daily practice”. Yet AI-driven tools like Google’s Gemini-powered live translation offering announced in August, alongside Apple’s equivalent with the AirPods Pro 3 rollout further dampens necessity-based learning, threatening Duolingo’s ability to retain subscriptions and drive revenue growth.

Freemium Model at Risk

Apple’s free live translation feature for its massive AirPods installed base also undercuts Duolingo’s freemium model. As necessity-driven learners shift to instant translation, Duolingo’s growth will hinge on leisure learners who are less likely to convert to paid subscriptions. While Duolingo monetizes its free leisure learners through ads and in-app purchases, these revenue streams grow far slower with less favourable margins than subscriptions. With 71% of Duolingo’s revenue generated from iOS users, Apple’s free live translation tool coupled with its ecosystem dominance will only distract leisure learners from Duolingo lessons, further eroding Duolingo’s DAUs and revenue growth.

This is consistent with management’s guidance for 2H25, which projects decelerating revenue growth of 35.5% in Q3 and 31.1% in Q4. Since the guidance was issued before Apple’s announcement of the live translation feature, Duolingo’s tempered outlook – which already signals DAU and subscription deceleration – now faces greater challenges that could result in an earnings shortfall and a deeper valuation reset.

Quantifying the Valuation Reset

Duolingo’s DAU and subscription revenue growth have been in steady deceleration in the last 12 months. With AI-driven competition intensifying, I forecast revenue growth at a five-year CAGR of 21.7%, reflecting ongoing pressure on necessity-based language learning subscriptions.

Meanwhile, margin improvements are expected, aided by Duolingo’s reduced exposure to Apple App Store commissions in the U.S. Although management expects external purchase flows to reduce the current 30% commission charged by Apple on U.S. iOS subscriptions to only 2% charged by Stripe for payment processing, the actual cost savings are likely to be less impactful. This is because iOS users – the bulk of Duolingo subscribers – still prefer in-app purchases facilitated within the Apple ecosystem for convenience and security.

A discounted cash flow (“DCF”) analysis yields a base case price target of $211 per share, assuming a 3.5% perpetual growth rate tied to long-term economic expansion in Duolingo’s core North American market.

In the downside scenario, where AI disruption further erodes necessity-based learning critical for Duolingo’s subscriptions, the company’s revenue growth could slow to a 17% five-year CAGR. This could push the stock to $162 per share – a level last seen in November 2023. While Duolingo’s recent diversification into chess offers some buffer, the discretionary offering potentially lacks the stickiness of language subscriptions, and does little to reduce the risk of AI-driven churn.

Conclusion

Duolingo’s valuation remains vulnerable to AI-driven disruption, with Apple’s AirPods Pro 3 live translation feature serving as a potent downside catalyst. By enabling seamless multi-lingual conversations, the free iOS-integrated tool threatens the necessity-driven learning that drives 71% of Duolingo’s current revenues.

With more than 500 million AirPods users likely to adopt the upcoming live translation update amid a robust upgrade cycle, Duolingo’s slowing DAU growth and freemium model face mounting pressure. And Duolingo’s recent diversification into chess lessons is unlikely to offset the extent of impending weakness in its core language business. As a result, Apple’s latest live translation feature is poised to weaken Duolingo’s long-term user retention and subscription growth, deepening risks of a valuation reset.

Disclaimer: This analysis is for informational purposes only and represents the opinions of Livy Research. It is not investment advice nor a recommendation to buy or sell the securities discussed.

Navitas Semiconductor (“Navitas”) has announced the abrupt resignation of founder and CEO Gene Sheridan on August 25. Industry veteran Chris Allexandre has been appointed to step in as the company’s new President and CEO effective September 1.

The leadership change comes at a critical juncture for Navitas, as management has just recently confirmed in its Q2 2025 earnings update a strategic pivot towards higher margin AI applications. Yet the latest CEO transition does not offer immediate resolution to Navitas’ mounting losses, which is at risk of being worsened by its imminent exposure to the 100% semiconductor tariffs tied to its manufacturing partnership with Taiwan’s Powerchip Semiconductor Manufacturing Corporation (“PSMC”) starting 1H26. As a result, my short thesis on Navitas – centered on the stock's overvaluation driven by its deteriorating margins and tariff-related cost pressures – remains intact despite the CEO transition news.

However, Allexandre’s extensive industry connections and expertise in power semiconductors does introduce long-term upside risks that warrant caution. In the below follow-up, I will evaluate the implications of Allexandre’s appointment to the role of Navitas CEO, highlighting potential opportunities that could challenge my bearish outlook on the stock while reaffirming near-term pressures from tariffs and fundamental weaknesses.

Recap of the Short Thesis

Navitas, a fabless Gallium nitride (“GaN”) and silicon carbide (“SiC”) power solutions provider, has seen its stock surge close to 80% year-to-date. It’s now trading at a lofty 32.4x 2025 sales following its non-exclusive partnership forged with Nvidia for the 800 Volt Data Center (“800 VDC”) architecture starting in 2027. This premium valuation far exceeds peers like Alpha and Omega Semiconductor (1.2x NTM P/S) and Monolithic Power Systems (13.7x NTM P/S), and ignores a critical downside catalyst – namely, Navitas’ impending exposure to 100% semiconductor tariffs on GaN wafers produced by Taiwan’s PSMC beginning 1H26.

With at least 11% of Navitas’ projected 2026 revenues (~$6.3 million of $57 million) tied to U.S. sales, these tariffs could add $6.3 million to its annual costs. This could erase the entirety of Navitas’ estimated $2 million U.S. gross profit next year (based on 37% GAAP gross margin in Q2 2025) and deepen its current losses.

The company also faces ongoing operational challenges. This includes expectations for steep revenue declines in 2H25, and a mere 1.6-year liquidity runway that’s poised to contract given Navitas’ upcoming exposure to additional semiconductor tariffs. Intensifying competition from larger integrated device manufacturers (“IDMs”) like Infineon and STMicroelectronics, coupled with protracted macro headwinds facing its core EV, solar and industrial end-markets, also threatens Navitas’ pricing power and pipeline conversion long-term.

As a result, the company remains exposed to threats of a 50% to 90% downside valuation reset. A correction to 13.7x NTM P/S (aligned with MPS’ profit-driven premium in the power semis peer group) implies a 50% downside to ~$3.10. A 7x NTM P/S (peer average, Navitas’ pre-Nvidia valuation) suggests a 74% downside to $1.60. In a worst-case scenario, mirroring unprofitable power semiconductor peers AOSL (1.2x NTM P/S) or Wolfspeed (1x NTM P/S), Navitas risks facing more than 90% downside to $0.27-$0.23, driven by its ballooning losses and dilution risks.

Allexandre’s Appointment: A Double-Edged Sword

Allexandre brings almost 28 years of sales and power semiconductor experience to Navitas, backed by his leadership roles at Renesas, NXP, Fairchild and Texas Instruments. His latest tenure as SVP and General Manager of Renesas’ Power division, alongside his oversight of the unit’s 2024 acquisition of GaN manufacturer Transphorm, aligns particularly with Navitas’ strategic penetration into AI opportunities long-term. Allexandre’s Board roles in TenXer Labs, Cyient Semiconductors, and Ambient Scientific also highlights a supportive network that could potentially further the application of Navitas’ GaN and SiC solutions across key AI, IoT and electrification verticals. While these credentials introduce opportunities for Navitas, they also pose risks to my short thesis on the stock, which necessitates a closer examination.

Implications for the Navitas Short Thesis

1. Persistent Tariff Exposure: The PSMC partnership remains Navitas’ Achilles’ heel, as U.S.-bound GaN wafers from the Taiwanese fab starting 1H26 will be subject to 100% semiconductor tariffs. Allexandre’s appointment does not immediately address this incremental cost headwind, which could deter prospective U.S. customers in AI verticals, reinforcing my thesis of imminent growth deteriorating and margin compression. This is poised to compound Navitas’ weak 2H25 revenue outlook and exacerbate the company’s ongoing losses, suggesting near-term financial strain that Allexandre’s appointment is unlikely to resolve.

2. Leadership Transition Uncertainty: Sheridan’s sudden and complete exit from the company – including the relinquishment of his Board seat – also raises concerns. Although he will stay with the company for another year to aid Allexandre’s transition into the CEO role, his abrupt departure and steep stock selling since the Nvidia partnership announcement in May signals potential misalignment with shareholder interests. As I’ve mentioned before – if Navitas is really in the early innings of AI infrastructure growth as management intends, then the substantial insider share selling (not just by Sheridan) since the Navitas-Nvidia collaboration announcement questions confidence in the opportunity’s scale.

Allexandre’s sudden appointment to the helm also introduces additional execution risks. Although his credentials and experiences are strong and relevant to Navitas’ pivot towards AI opportunities, his strategic vision for the company – which remains uncertain – could exacerbate near-term growth challenges, supporting my bearish outlook on the stock.

3. Strategic Pivot Risks: The Board’s optimism about Allexandre’s “track record of driving transformation and delivering sustainable and profitable growth, operational excellence and business leadership in power semiconductor markets” also overlooks the capital-intensive nature of Navitas’ AI pivot.

Allexandre was likely hired to further Navitas’ focus on emerging AI opportunities – a costly strategic pivot that’s already impacting its near-term growth profile. As I’ve discussed in the initial analysis, the company’s $161.2 million cash balance only provides about 1.6 years of liquidity runway, which is poised to contract with the incremental semiconductor tariffs. This means Allexandre’s first order of business would still be to improve Navitas’ AI-driven backlog and diversify its manufacturing.

But these endeavours will take time to materialize, triggering additional financing needs for the company before its AI-designated productions start to ramp in later 2026. And it’s unlikely the company will tap into debt financing, as it’d increase its exposure to additional interest expenses and, inadvertently, further margin erosion. This implies another equity issuance is potentially on the way. Although this could represent a prudent capitalization of the stock’s current valuation premium, it’d come at the expense of inevitable share dilution for investors.

Potential Pros of Allexandre’s Appointment

Although Allexandre’s appointment doesn’t immediately mitigate Navitas’ imminent tariff exposure nor losses, his background still offers long-term upside scenarios that could challenge my short thesis:

1. Increased Deal Wins in AI Verticals: Allexandre brings almost three decades of sales experience in the semiconductor industry, as evidenced in his past tenures at Texas Instrument, NXP, Fairchild, Integrated Device Technology (“IDT”, acquired by Renesas in 2019), and Renesas. This relevance to Navitas is further reinforced by his latest role as the Power division head at Renesas, where he’s led penetration into data center infrastructure opportunities. The background likely encompasses a deep network of potential customers and opportunities for Navitas, as it looks to establish a stronger foothold in the AI vertical.

Allexandre’s participation in Renesas’s 2024 acquisition of Transphorm also implies expertise in power semiconductor solutions supply chain management. This represents critical experience in leading Navitas’ diversification of its GaN wafer manufacturing footprint.

Taken together, Allexandre’s expertise could help strengthen Navitas’ AI deal wins, driving markets to overlook the company’s near-term exposure to tariffs and ballooning losses. However, all of these potential upsides remain speculative and may not materialize until 2027 when Navitas’ AI products scale. This accordingly leaves the immediate tariff overhang on the stock’s current valuation premium unaddressed.

2. Potential Renesas Partnership for U.S.-Based Manufacturing: Allexandre’s deep ties to Renesas could open doors to a potential partnership and/or product collaboration with Navitas.

Renesas currently has U.S.-based GaN/SiC wafer manufacturing capacity via its partnership with Wolfspeed and Polar Semiconductor. The company also owns U.S.-headquartered Transphorm, though the subsidiary’s main GaN wafer manufacturing facility is currently located in Japan. But because of emerging opportunities for GaN/SiC application in next-generation AI infrastructure, it’s unlikely Renesas would be willing to contract its U.S.-based capacity out to Navitas in the near-term. This is consistent with the fact that Renesas has also been named a key silicon provider for Nvidia’s next-generation 800 VDC architecture, which directly rivals Navitas.

However, in the longer-term, there is potential for R&D collaboration between the two companies that’d allow Navitas to diversify its GaN manufacturing footprint, and mitigate its exposure to semiconductor tariffs. This would be consistent with the partnership observed between Renesas and TenXer Labs – where Allexandre’s recently been appointed as a Board member – since 2021.

3. Buyout Potential: Sheridan’s sudden exit and Allexandre’s appointment also unlocks speculation of a potential Navitas sale. And Renesas could be a relevant suitor, given Allexandre’s long-time ties with the company.

Renesas, with its $1.5 billion cash balance and sustained profit accretion, could easily acquire Navitas at its current $1.3 billion market cap. And there’s incentive to do so, as such an acquisition could bolster Renesas’ GaN/SiC portfolio for AI applications over the longer-term. Such a buyout could also help Navitas diversify its current GaN wafer manufacturing capacity away from PSMC, and yield immediate cost synergies for Renesas. And Allexandre’s familiarity with Renesas’ operations and his previous role in the Transphorm acquisition brings relevant expertise to bridge such a deal.

Although there’s yet to be any public indications of such a scenario taking place, especially as Navitas’ Board emphasizes focus on independent growth under Allexandre’s leadership, it shouldn’t be overlooked. A Navitas buyout would cap immediate downside risks facing the stock and undermine my short thesis.

Conclusion: Short Thesis Intact, But Upside Risks Emerge

Navitas’ recent appointment of Allexandre to the CEO role risks countering my short thesis on the stock. His background is of extreme relevance to Navitas’ strategic pivot towards higher-margin AI applications in the long-term. For instance, Allexandre’s decades of sales experience and network acquired could reinforce Navitas’ penetration into emerging AI infrastructure opportunities. Meanwhile, his industry expertise acquired from leading Renesas’ Power division could help Navitas diversify its supply chain, improve its fundamentals, and even pave way for a potential buyout that could unlock shareholder value.

However, the core of my short thesis – namely, Navitas’ exposure to 100% semiconductor tariffs and persistent losses – remains unaddressed with Allexandre’s appointment. Coupled with a weak 2H25 outlook, a compressed liquidity runway, and intensifying competition, Navitas faces a 50% to 90% downside risk. As a result, I maintain my short thesis on the stock, but advise caution as impending clarity on Allexandre’s AI vision and business plan for Navitas could spark volatility – especially if markets embrace his strategic initiatives.

Disclaimer: This analysis is for informational purposes only and represents the opinions of Livy Research. It is not investment advice nor a recommendation to buy or sell the securities discussed.

Navitas Semiconductor Corporation (“Navitas”), maker of gallium nitride (“GaN”) and silicon carbide (“SiC”) power solutions, has been riding high on investor optimism in recent months. The stock has been trading at 28.8x 2025 sales since announcing its non-exclusive collaboration with Nvidia to supply power solutions for the 800 Volt Data Center (“800 VDC”) architecture starting 2027 – a sizable jump from about 7x prior. Its current valuation significantly outpaces comparable peers like Alpha and Omega Semiconductor (“AOSL”) at 1.6x NTM P/S, and even profitable peer Monolithic Power Systems (“MPS”) at 11.6x NTM P/S.

However, this premium valuation fails to price in a critical catalyst – namely, 100% semiconductor tariffs on GaN wafers produced at Taiwan’s Powerchip Semiconductor Manufacturing Corporation (“PSMC”) starting 1H26. With Navitas’ 11% U.S. revenue mix ($6.3 million of $57 million projected for 2026) exposed to these tariffs, the resulting cost increases threaten to deepen losses, accelerate cash burn, and trigger dilutive financing. This risks a valuation reset akin to power semiconductor peer Wolfspeed’s collapse from 10x to 1x NTM P/S post-Chapter 11 bankruptcy filed in June. Intensifying competition and macro headwinds risk further amplifying this downside, making Navitas’ price of $6.60 as of August 11 close a compelling short opportunity.

100% Semiconductor Tariffs: The Catalyst for Re-Rate

While Navitas doesn’t bifurcate its revenue disclosure by product, management commentary in the past has provided clarification that the bulk of the company’s sales are generated from its GaN portfolio. Meanwhile, Navitas’ SiC products currently represent only a nominal share of its revenue mix, with the majority generated from sales to China.

While SiC represents a minority of our total revenue, our products are produced in the U.S. at [X-FAB] and the majority is sold in China…We anticipate GaN revenues, which represent the significant majority of our revenue today, to have a very limited direct impact from tariffs as our GaN products are manufactured in Taiwan and are sold predominantly outside the U.S. – Navitas Q1 2025 Earnings Call Transcript

TSMC, Navitas’ current sole GaN wafer supplier, has announced it will cease GaN production on July 31, 2027. This has prompted Navitas to transition its GaN wafer production to PSMC starting 1H26. The company will first start with ramping 80-200V GaN at PSMC’s 8-inch Fab 8B in Taiwan, then introduce its higher voltage GaN products to the mix over the next two years. Unlike TSMC, which has committed significant investments into building out its U.S. manufacturing capacity, PSMC’s capacity remains limited to Taiwan. This accordingly makes its GaN exports to the U.S. vulnerable to 100% semiconductor tariffs recently proposed under the Trump administration. With at least 11% of Navitas’ 2026 revenue (~$6.3 million) tied to U.S. sales – potentially more when GaN supplies to Nvidia start to ramp – these tariffs could add $6.3 million to the company’s annual costs.

At about 37% GAAP gross margin in Q2, excluding the one-time $3 million SiC inventory reserve, U.S. gross profit estimated for 2026 could reach $2 million. But tariffs risk erasing this gain by introducing an additional $6.3 million cost headwind. Not only is the set-up expected to worsen Navitas’ losses, but passing on the additional costs to customers could risk demand erosion, as competition limits pricing power (further discussed in later sections).

Although Navitas’ 60% Hong Kong revenue mix mitigates some exposure, the company’s current tariff headwinds faced in China could be proxy for upcoming challenges to its fundamental outlook if the 100% semiconductor duties are applied to U.S. sales. Navitas currently generates about 12% of its revenue from China, generated primarily from SiC sales. Since Navitas currently produces its SiC wafers at X-FAB in the U.S., related shipments to China have been subject to the region’s 10% tariffs. This has caused management to take a $3 million reserve on SiC inventory in Q2, accompanied by a projected 50% y/y decline in Q3 revenue to $10 million due to soft demand in China EV and industrial markets as “customers wait for improved economic indicators”.

And implications from the 100% tariffs on PSMC GaN wafers imported from Taiwan starting next year will likely be worse. Although Navitas is also “evaluating additional suppliers to enhance supply chain resilience”, which could include U.S.-based manufacturers, PSMC’s scale likely makes it the company’s primary GaN production partner in the foreseeable future. This is consistent with Navitas’ increasing focus on higher voltage GaN solutions, such as its GaNSafe ICs that currently require 200mm technology. GlobalFoundries is currently the only U.S.-based foundry offering external volumes for 200mm GaN wafer manufacturing at scale in Vermont. But since this capacity is still in early ramp-up, Navitas will likely need to rely on PSMC for the majority of its supply in the upcoming transition to AI data center opportunities. This accordingly locks in incremental tariff risks that could trigger a valuation re-rate by 1H26 for Navitas.

Valuation Disconnect: Premium Multiple, Weak Fundamentals

At 28.8x 2025 sales, Navitas’ valuation today lacks constructive support from its underlying business fundamentals. The multiple also far exceeds its power semiconductor peers. AOSL, which is also an unprofitable fabless GaN solutions provider, currently trades at 1.6x NTM P/S. Even Monolithic Power Systems (“MPS”), which is currently profitable, is only trading at 11.6x NTM P/S. More importantly, Wolfspeed is a cautionary tale in the industry, with its valuation falling from 10x NTM P/S just a year ago to 1x following its Chapter 11 bankruptcy filing in June, highlighting the extent of downside risks facing unprofitable GaN/SiC constituents.

Navitas’ ballooning losses to $49 million in Q2 and fading prospects of 2026 EBITDA breakeven – a target that was reiterated by management during the Q1 earnings update and absent in Q2 – also signals a weak fundamental outlook. This is exacerbated by a sharp reversal in management’s earlier optimism for 2H25 growth acceleration based on $450 million in new design wins in 2024 and a $2.4 billion customer pipeline. This was evidenced in management’s guidance for a 50% y/y decline in Q3 revenue to $10 million, reflecting delays in pipeline conversion due to tariff uncertainties in China. Yet the premium allocated to Navitas’ valuation at current levels continues to ignore these risks – especially with its additional exposure to the impending 100% semiconductor tariffs, which sets the stage for an imminent re-rate.

Cash Burn and Dilution: Tariffs Fuel a Vicious Cycle

Navitas’ current $161.2 million cash balance is accompanied by a $25 million quarterly operating cash burn run-rate, implying a 1.6-year liquidity runway. And its exposure to additional semiconductor tariffs starting next year could potentially increase this cash burn by $6.3 million to $106 million, shrinking its liquidity runway to under 1.5 years. This could necessitate further financing before the early production ramp of its AI-designated productions starting in later 2026.

It's likely Navitas will be incentivized to capitalize on its valuation premium at current levels and issue new shares to raise additional capital. The company had disclosed it’s already raised $97 million in net proceeds by issuing 20 million shares through two separate $50 million-increment at-the-market programs in a recently filed press release. The share sale had triggered a 22% drop in the stock during the week it was executed. Since tariffs are expected to deepen its cash needs, Navitas could potentially tap into another ATM offering. A similar $50 million increment issuance could add 7 to 8 million shares at current prices ($6.60 August 11 close), implying dilution of about 4% alongside additional EPS compression. This accordingly introduces a vicious downward spiral for Navitas – cost increases are expected to erode margins, forcing dilutive share issuances, which will depress the stock’s valuation premium multiple and compress EPS metrics, further fuelling losses and liquidity needs.

Competition: Non-Exclusive Nvidia Deal Limits Upside

Navitas’ valuation premium at 28.8x 2025 sales also overlooks the non-exclusive nature of its collaboration with Nvidia on the next-generation 800 VDC architecture, with competition having only intensified since the initial disclosure in May.

In Nvidia’s original press release in May, it had listed five other silicon providers in addition to Navitas for its 800 VDC architecture.

They included Infineon, MPS, ROHM, STMicroelectronics and Texas Instruments. This list was quietly expanded in late July to include four additional contenders – Analog Devices, Innoscience, OnSemi and Renesas – without an amendment date to the original May 20 press release.

Of these 10 silicon suppliers, eight (excluding ROHM and Analog Devices) have GaN/SiC expertise for AI data center applications. Most of them are IDMs, with some adopting a hybrid manufacturing strategy that includes third-party foundry capacity. These competitors, with their larger revenue bases and R&D budgets, could threaten Navitas’ deal size potential with Nvidia, and limit the company’s pricing power and ensuing margin expansion prospects.

Navitas’ technical moat demonstrated through its GaNSafe ICs, which feature 350ns short-circuit protection and 2kV ESD shielding critical for high-power AI data center applications, is also at risk of being overtaken by competition. The embedded short-circuit protection in Navitas’ GaNSafe ICs is critical in protecting systems under faulty conditions and reducing failure risks. GaNSafe’s 2kV electrostatic discharge (“ESD”) shielding also improves protection and reliability of devices under harsh environments across automotive, industrial and AI data center applications.

However, larger and deeper pocket competitors like Infineon and STMicroelectronics, which are also suppliers of Nvidia’s 800 VDC architecture, are rapidly narrowing this R&D gap. Infineon’s 300mm GaN wafer production – industry’s first – is expected to boast superior yields and additional cost efficiencies, while its U.S.-based manufacturing also mitigates tariff risks. Meanwhile, STMicroelectronics is also advancing integration capabilities in its Sti²GaN product family, underscoring potential catch-up to GaNSafe’s protection features by 2027 when AI data center applications start to ramp.

The mounting competitive pressure against Navitas now risks undermining its potential share of Nvidia’s 800 VDC architecture opportunities. The 100% semiconductor tariff risk on imported PSMC GaN wafers could further limit the cost efficiency of Navitas’ GaNSafe ICs, diluting its appeal against rival offerings that benefit from U.S.-based manufacturing at scale.

Macro Headwinds: Amplifying Tariff Impact

The company’s also been experiencing protracted cyclical pressures across its core EV, solar and industrial end-markets over the past year. This has been further challenged by broad-based U.S.-China trade tensions, with tariff uncertainties potentially curtailing new design wins and delaying related conversions. This is consistent with management’s recent commentary that its China EV and industrial customers are currently awaiting “improved economic indicators”, which limits visibility into its pipeline conversion outlook. The upcoming removal of solar and EV tax credits in September under Trump’s “One Big Beautiful Bill” (“OBBB”) is likely to further obscure recovery prospects across Navitas’ core growth verticals.

Not only does this cloud Navitas’ growth outlook, but it also compounds near-term tariffs in impacting the company’s profit trajectory. Coupled with management’s intensions to “continue investing in next-generation GaN and SiC technology platforms to serve the increasing power consumption across AI data center and energy infrastructure markets”, which entails elevated R&D, the company faces imminent margin deterioration. It’s likely that management’s guidance for EBITDA breakeven by 2026 will be officially retracted, subjecting Navitas’ current valuation at 28.8x 2025 sales to imminent risks of a downward re-rate. Management’s omission of the 2026 EBITDA breakeven target during the Q2 earnings call was likely a cautionary tale.

Valuation Reset: A 44–90% Downside

Navitas’ 28.8x 2025 P/S is unsustainable given tariff risks, extended losses, and intensifying competitive pressures. A correction to 14.5x NTM P/S (aligned with MPS’ profitable growth) implies a 44% downside to ~$3.66. A 7x NTM P/S (peer average, Navitas’ pre-NVIDIA valuation) suggests a 73% downside to $1.82. In a worst-case scenario, mirroring AOSL (1.6x NTM P/S) or Wolfspeed (1x NTM P/S), Navitas could face more than 90% downside to ~$0.42–$0.67, driven by ballooning losses and dilution. Despite the $450 million in new 2024 deal wins and the upcoming Nvidia collaboration, near-term risks outweigh long-term upside at current valuations.

Conclusion

Navitas’ premium valuation ignores a myriad of imminent fundamental risks: 100% U.S. tariffs on PSMC GaN, a 1.6-year cash runway vulnerable to dilution, intensifying competition from larger IDMs, and macro headwinds delaying pipeline conversion. The lack of immediate disclosure pertaining to its material exposure to the upcoming 100% semiconductor tariffs, alongside substantial insider share selling since the announcement of Navitas’ collaboration with Nvidia, also suggests a lack of management prioritization over shareholder interests. With Navitas’ valuation premium formed on weak fundamentals, its risk-reward is evidently scaled to the downside – with its exposure to 100% semiconductor tariffs and intensifying competition serving as an imminent catalysts for correction.

Disclaimer: This analysis is for informational purposes only and represents the opinions of Livy Research. It is not investment advice nor a recommendation to buy or sell the securities discussed.

Nvidia’s valuation continues to hinge on robust investor sentiment over the AI spending environment as it forges new all-time highs. But a critical oversight may be superficially skewing market expectations. The following analysis will examine the underappreciated implications of ASC 842 Lease Accounting, which yields two material risks to Nvidia’s growth and valuation outlook. First, it highlights how hyperscaler capex – a key barometer of Nvidia demand – is increasingly tied not only to long-term leases on structural components like office buildings and data centers, but also GPUs. This dynamic raises embedded risks of double-counting AI GPU demand, and underscores a potential overstatement of $11 billion in Nvidia’s anticipated data center sales for FY 2026 alone. Second, the analysis examines fundamental vulnerabilities in CoreWeave’s GPU depreciation and capex strategy. It’ll dive into finance lease dynamics observed across its key hyperscaler customers – particularly, Microsoft. The findings corroborate escalated risks of imminent asset impairment, which Nvidia is exposed to given its 7% stake in CoreWeave.

Introduction

Hyperscaler capex spend has been a key focus area in recent years, as it serves as a critical gauge for impending AI infrastructure commitments – and, most importantly, a direct barometer for Nvidia’s demand outlook. The growing influence of AI-related capex trends on both valuation and market sentiment was evident earlier this year, when training efficiency revelations from DeepSeek spooked investors about potential spending fatigue in the industry. At the time, markets witnessed Nvidia losing 17% of its value in a single day, wiping out about $590 billion from its market cap to mark the worst intra-day decline since March 2020. Subsequent headlines about Microsoft potentially pulling out of planned data center leases only further amplified market anxiety. It wasn’t until the U.S. “Big 4” hyperscalers (i.e., Microsoft, Google, Amazon and Meta Platforms) reaffirmed their full-year AI spending commitments during the March quarter earnings season that assuaged investor fears.

While strong headline capex commitment from the Big 4 U.S. hyperscalers have fueled optimism around Nvidia’s growth trajectory, there has been relatively limited scrutiny into how much of that spending really flows directly to chipmakers. And this isn’t merely about the allocation between AI processors vs. other long-lived assets (e.g., fulfilment facilities and equipment, office buildings, data center colocations, etc.), or merchant GPUs vs. custom ASICs, or even Nvidia vs. its competing rivals.

What remains largely overlooked is the distinction between capex allocated to direct hardware purchases vs. leases. Specifically, the implications of ASC 842 Lease Accounting warrant greater attention, given rising relevance of long-term leases amid surging AI infrastructure investments and the emergence of the GPU rental market.

As evidenced in neo-cloud operator CoreWeave’s sustained rally since its IPO earlier this year and management’s emphasis on surging demand for its core GPU rental business, the capital deployment landscape is rapidly evolving. Long-term leases are no longer a nominal component in hyperscaler capex commitments that’ve historically been tied to structural components of operations like office buildings, data center colocation facilities, and/or equipment. Instead, they now represent a strategic avenue for accessing high-value AI infrastructure, particularly GPUs, critical to industry growth roadmaps.

This shift has accordingly levied material implications for how hyperscaler capex should be interpreted moving forward – both in terms of spend visibility and its direct impact on AI technology enablers like Nvidia. In this context, understanding ASC 842 Lease Accounting is no longer optional. Only through a clear grasp of how finance leases are reported under ASC 842 can investors accurately assess Nvidia’s near-term monetization potential and avoid overestimating demand embedded in headline capex figures.

In the following analysis:

I’ll begin with a brief primer on ASC 842 Lease Accounting.

I’ll then proceed with a two-part analysis that reveals two underappreciated, yet imminent and material risks to Nvidia’s growth outlook that stem from these accounting implications:

1) The overstatement in growth expectations currently underpinning Nvidia’s rally to all-time highs; and

2) The growing mismatch between hyperscaler GPU rental demand and CoreWeave’s expansion plans – a dynamic that poses indirect risks to Nvidia, given its role as a strategic investor in the neo-cloud operator.

What is ASC 842 Lease Accounting?

ASC 842 Leases is the U.S. GAAP standard employed by many reporting participants across the AI value chain today, including the Big 4 U.S. hyperscalers, Nvidia, and CoreWeave. The guidance provides oversight on leases of all long-lived assets (e.g., PP&E), including the differentiation between short- and long-term leases, and how each type is to be accounted for in financial reporting.

Short-Term Leases

Under ASC 842, a short-term lease is defined as “a lease that has a lease term of 12 months or less and does not include a purchase option that the lessee is reasonably certain to exercise”. For instance, if an enterprise customer goes to a neo-cloud operator like CoreWeave or Nebius, and rents a couple of hours of H100 capacity under a 1-month contract, then it’s a short-term lease.

Long-Term Leases

**The following may be quite confusing for the day-to-day investor. If you’re not interested, and just want to know enough to understand the imminent risks facing Nvidia and CoreWeave, then here’s the condensed version:

There are two types of long-term leases: 1) finance lease, and 2) operating lease. A finance lease would represent a lease where the lessee assumes ownership of the underlying asset; the contract value over the lease term is equal to or exceeds the fair value of the underlying asset; and/or the lease term is for the major part of the remaining economic life of the underlying asset. For instance, if a hyperscaler rents GPU capacity from a neo-cloud operator for five years out of the six-year remaining useful life designated for the underlying GPU, then the arrangement would be classified as a finance lease.

Alternatively, any long-term lease exceeding 12 months that does not exhibit any of the aforementioned characteristics would be classified as an operating lease. For instance, if a hyperscaler rents GPU capacity from a neo-cloud operator for two out of the six-year remaining useful life designated for the underlying GPU, and the total lease payment only represents the minority of the underlying GPU’s fair value, then the arrangement would be classified as an operating lease.**

Now the real deal on classification, initial measurement and subsequent accounting for long-term leases is detailed as follows:

Under ASC 842, long-term leases are defined as any lease that has a lease term of more than 12 months. There are two types of classifications for long-term leases:

Finance lease: Any long-term lease with a lease term of more than 12 months and exhibits one or more of the following characteristics:

o The lease transfers ownership of the underlying asset to the lessee

o The lease includes an option for the lessee to purchase the underlying asset that the lessee is reasonably certain to exercise

o The lease term is for the major part of the remaining economic life of the underlying asset

o The present value of the sum of lease payments over the lease term and any residual value guaranteed by the lessee is equal to or exceeds substantially all of the fair value of the underlying asset

o The underlying asset is specially curated for the lessee and has no alternative use to the lessor

Operating lease: Any long-term lease with a lease term of more than 12 months that does not exhibit any of the criteria required for finance lease accounting.

Every lessee will need to perform a classification assessment of the lease as of commencement date, as that would dictate the subsequent accounting and financial reporting disclosures for the arrangement.

Regardless of whether the long-term lease is classified as a finance or operating lease, the underlying asset involved would be considered a “right-of-use” – or “ROU” – asset with a matching lease liability that’s recorded on the lessee’s balance sheet.

The lease liability recorded at lease commencement should reflect the present value of total lease payments over the lease term. And the ROU asset recorded at lease commencement would represent the total of the initial measurement of the lease liability; any additional amounts made to the lessor related to the lease at or before the commencement date, net of lease incentives; and any indirect costs incurred by the lessee (e.g., commissions). If there are no additional amounts made to the lessor at lease commencement and no indirect costs involved, then the ROU asset should match the lease liability at initial measurement and reporting on the lessee’s balance sheet.

The subsequent wind-off of lease liability and ROU asset amortization is as follows:

ROU asset: Amortized using the straight-line method over the shorter of the useful life of the asset or the lease term. For instance, if an enterprise customer enters into a long-term operating or finance lease for GPU capacity from a hyperscaler or neo-cloud over a two-year lease term, but the underlying GPU asset has a remaining useful life of six years, then the enterprise customers will amortize the ROU asset over two years.

Lease liability: The lease liability is reduced annually based on the difference between the annual lease payment and the annual interest expense accrued on the total lease liability outstanding. This means the lease liability reduction will be smaller during the earlier years of the lease term (due to greater interest accrued) and increase over time.

It’s the subsequent impact on the income statement that differs between finance lease and operating lease accounting:

Finance lease: An annual interest expense accrued on the outstanding lease liability is recorded on the income statement. Because the outstanding lease liability is greater during the earlier years of the lease term, the annual interest expense impact on the income statement is also accordingly greater. This results in a frontloaded cost impact on the income statement for finance leases, which declines over time.

Operating lease: In contrast, operating lease expense on the income statement is recorded on a straight-line basis over the lease term. If the annual lease payment is $10,000, then the same amount will be recorded as an operating lease expense on the income statement. Although the accrued interest is still considered in the annual lease liability wind-off exercise, it’s not separately reported on the income statement.

Part I: Market Estimates for Nvidia GPU Demand are Overstated

During Nvidia’s Q4 FY 2025 earnings call, CFO Colette Kress disclosed that the U.S.’ largest hyperscalers now account for more than half of the company’s data center revenue, with CEO Jensen Huang adding that demand stemming from the cohort has primarily been driven by internal consumption needs:

“In Q4, large CSPs represented about half of our data center revenue…And the CSPs have internal consumption and external consumption, as you say. And we are using – of course, used for internal consumption.” - Nvidia Q4 FY 2025 Earnings Call

This has accordingly made capex guidance from the Big 4 U.S. hyperscalers a key barometer for Nvidia’s forward demand. To date, Microsoft, Google, Amazon and Meta have collectively announced more than $310 billion of capex for the current year, with potential for further increase towards more than $330 billion in 2026. The bulk have been attributable to AI infrastructure buildouts, including servers equipped with Nvidia GPUs.

These figures have also become key inputs in Nvidia’s growth forecasts and stock price projections among top Wall Street analysts:

Yet based on the compilation of institutional equity analyst projections and analyses, none have really paid close attention to amounts attributable to finance leases, which are commonly disclosed by management as part of their capex outlook.

As discussed in the above crash course on ASC 842 lease accounting, a finance lease relates to one where the lessee – or hyperscaler, in this case – essentially assumes ownership of the underlying asset. This is achieved either by committed use of the underlying asset over the majority of its remaining useful life or by extracting the majority of the underlying asset’s value over the lease term. As much as 20% of hyperscaler capex commitments in 2024 have been attributable to finance leases, up from 15% in 2023. Total finance lease capex incurred by the U.S. Big 4 hyperscalers in 2024 have increased by 27% from the prior year:

The consistent increase in finance lease capex aligns with hyperscalers’ aggressive expansion of their data center and colocation footprint in recent years to facilitate burgeoning AI demand. In addition to related land and building leases, another key contributor to the increase in finance lease capex is GPU rentals – an emerging corner in the AI value chain that’s only gained relevance in the last two years.

The GPU rental market is currently dominated by hyperscalers themselves, as well as a niche cohort of neo-cloud operators, including CoreWeave and Nebius. CoreWeave generated close to $1 billion in revenue during Q1 2025 alone – four times higher than the same quarter last year and already accounting for nearly half of its total $1.9 billion revenue from all of 2024. Nebius showed similar momentum, with Q1 2025 sales up four-fold y/y, also representing about half of its 2024 revenue. Hyperscalers have been key demand drivers for the GPU rental market, as the lease-based model grants them access to immediate deployment, flexibility and cost efficiencies.

This explosive growth underscores the rising relevance of GPU rentals – and consequently, finance leases – in accelerating AI infrastructure deployment. Based on recent observations, it can be safely assumed that at least 7% of planned 2025 and 2026 capex amongst the U.S.’ largest hyperscalers will be represented by finance lease arrangements, with about half linked to GPU rentals. That’s $11 billion in 2025 AI capex that will not go directly to merchant GPU purchases from Nvidia, but rather to neo-cloud providers like CoreWeave and Nebius instead.

And here’s the issue: In addition to capex from the U.S. Big 4 hyperscalers, markets are increasingly baking in spending from neo-cloud providers to support a more bullish outlook for Nvidia’s growth. This is reflected in analyst AI capex estimates for 2025 and 2026, which consistently exceed the guidance provided by the Big 4 alone. The overshoot suggests markets are factoring in incremental demand from other sources – such as other cloud-service providers like Oracle, enterprise and sovereign buyers, and neo-cloud operators like CoreWeave and Nebius.

Wedbush, for example, is one of the analysts that’ve expressed incremental optimism tied specifically to CoreWeave’s capex on top of what the U.S. Big 4 hyperscalers have committed to in inferring Nvidia’s strong demand outlook:

But what’s really happening is that markets are at risk of overstating Nvidia’s demand outlook by double-counting finance lease GPU commitments. As noted earlier, about $11 billion of the U.S. Big 4 hyperscalers’ planned capex is expected to be fulfilled through GPU rentals instead of direct purchases. That demand will ultimately flow through to Nvidia via neo-cloud providers like CoreWeave and Nebius, which have collectively planned $25 billion in 2025 capex.

As a result, AI hardware demand that Nvidia’s poised to benefit from would be the $310 billion from hyperscalers – and not an inflated sum that’s inclusive of neo-cloud operators’ $25 billion of planned capex spend.

Considering Nvidia is currently trading at 19x NTM sales, with markets at risk of mistakenly pricing in $11 billion in overstated GPU demand, the stock is potentially overvalued by $210 billion – or more than 5% of its market cap – today. Or to put into better perspective, a third of Nvidia’s $620 billion rally this year is likely attributable to overstated revenue expectations.

The broader impact on sentiment could pose an even greater headwind. A similar dynamic was evidenced in mid-April, when Nvidia shares dropped nearly 14% during the week that followed the U.S. government’s announcement of export curbs on the H20 GPU to China, alongside management’s subsequent disclosure of a related $5.5 billion write-off. While applying a 19x multiple to the write-off would imply a $105 billion impact, the sharper selloff likely reflected a broader market re-rating around Nvidia’s longer-term growth curve, which was lowered as a result of weakened penetration propsects into China AI opportunities.

A comparable correction could follow once markets fully recognize the extent to which current growth expectations priced into Nvidia’s valuation are overstated. The adjustment likely wouldn’t stop at the $11 billion overstatement estimated for the current year. Instead, it’d likely have a compounding effect by lowering Nvidia’s multi-year growth curve and, inadvertently, its valuation multiple as well.

More importantly, signs of intensifying competition are also emerging based on the consistent decline in hyperscaler capex (ex-finance leases) as a percentage of Nvidia data center revenue in recent years.

This observation potentially implies diversion of spend to competitors and/or other AI infrastructure components (e.g., software, server racks and infrastructure, etc.). For instance, Google has recently been giving greater focus on revenue-linked investments according to management’s commentary during the Q1 2025 earnings call, which infers increasing GPU deployments:

“And as you've seen in the comment I've just made on cloud, we do have demand that exceeds our available capacity, so we'll be working hard to address that and make sure we bring more capacity online.” - Google Q1 2025 Earnings Call

Yet the consistent decline in Google’s capex (ex-finance leases) as a percentage of Nvidia’s data center revenue suggests the GPU install mix is increasingly being diverted to rivals – like Broadcom, given Google’s heightened focus on its in-house developed TPU. This would also be consistent with Google’s increasingly vertically integrated strategy:

“We also look at every investment that we make to ensure that we're doing it in the most cost-efficient way to optimize our data center. As you know, our strategy is mostly to rely on our own self-design and build data centers. So they're industry-leading in terms of both cost and power efficiency at scale. We have our own customized TPUs. They're customized for our own workload. So, they do deliver outstanding superior performance and CapEx efficiency.” - Google Q4 2024 Earnings Call

Taken together, Nvidia’s all-time high valuation today appears to be priced for more than perfection. And it has certainly yet to de-risk for double-counting in estimates due to market’s continued underappreciation for the finance lease component of hyperscaler capex. This is further compounded by the growing diversion of related investments toward Nvidia’s rivals instead.

Part II: Mismatch in GPU Lease Demand vs. CoreWeave Capex and Depreciation Schedule

There’s also been much debate on whether CoreWeave has employed appropriate accounting on the valuation of its long-lived assets – particularly, the GPUs currently underpinning its core business model. The discrepancy between the six-year useful life CoreWeave has attributed to its GPUs versus the much shorter span employed by comparable peer Nebius, alongside the accelerating decline in GPU rental rates due to the rapid cadence of tech advancements has raised causes for caution.

Nebius has specifically disclosed in its 2024 Form 20-F that its GPU assets are depreciated using the straight-line method over a four-year useful life. And during its Q1 2025 earnings call, management had provided updated clarification that the company now uses a “full year depreciation schedule” for its GPU assets. While it’s unclear whether this suggests the newly acquired GPUs are fully depreciated within one year, or if they retain a four-year useful life but are depreciated in full during the year of purchase rather than pro-rated, either case implies a much shorter useful life than what CoreWeave’s employed. Even at the most generous estimate, Nebius is likely depreciating its GPU assets over an average of about three years. This would align with Nebius’ evenly distributed quarterly capex. In Q1 2025, the company spent $544 million in capex, just over a quarter of its $2 billion full-year guidance. This suggests roughly half of its planned GPU purchases would be back-end loaded, but subject to full-year depreciation, leaving about three years of remaining useful life.

As I’ve explained before, “useful life” is a strictly defined term in GAAP. The Accounting Standards Codification employed by U.S. GAAP defines useful life as “the period over which an asset is expected to contribute directly or indirectly to future cash flows” – or the period over which the asset is expected to generate an economic value to the business.

This means CoreWeave’s assignment of a six-year useful life to its GPU assets implies an expectation that the hardware will not only remain operational over that period, but also continue to generate economic value for the company throughout. But here’s the problem – Nvidia has recently committed to a one-year advancement cadence to its product roadmap. This has effectively accelerated the pace of hardware obsolescence, leading to a rapid decline in the price and rental rate of past-generation GPUs. With everyone demanding the latest and greatest, which promises better performance and TCO, rental rates on past-generation GPUs like the H100 have dropped significantly from its peak. In early June, Amazon disclosed it has slashed rental prices on older generations of Nvidia GPUs by as much as 45%:

CoreWeave itself, and peer Nebius, are also currently offering H100 rentals at as low as $2 per hour, which is a whopping 75% drop from peak prices at about $8 per hour two years ago, and 23% drop from just a little under a year ago. This accordingly highlights the adverse shift in business unit economics for CoreWeave, which raises doubt on whether its currently extended GPU depreciation schedule is appropriate.

With GPU rental rates facing a steepening decline as new advancements get introduced to market at a rapid pace, CoreWeave now faces a tighter timeline to breakeven. If a GPU investment made today isn’t fully recoverable within two years of the designated six-year useful life, then it’s poised for impairment given limited demand and price accretion expected in the remaining four years. U.S. GAAP currently requires reporting companies to assess the asset useful life and recoverable value annually for impairment. If the total recoverable amount of the asset is expected to fall below its cost recorded at initial measurement, then the difference will need to be written off from the balance sheet with an impairment expense.

This is a very real and material risk for CoreWeave that’s being underappreciated by markets, as evidenced in the stock’s continued upsurge to new all-time highs.

In CoreWeave’s latest earnings call, management sought to assuage these concerns by clarifying the company employs a “success-based” capex strategy, whereby investments are only deployed when there’s sufficient contracted revenue to ensure breakeven within the GPU’s useful life:

“Our capital expenditures are success based. Substantially, we enter into compute CapEx programs when we sign multiyear contracted revenue that more than covers the cost of the CapEx within the contract terms. This enables us to responsibly scale our debt structures that support this contractual revenue and utilize naturally deleveraging self-amortizing debt facilities that allow for us to maintain a relatively low leverage multiples.” - CoreWeave Q1 2025 Earnings Call

But the only way to mitigate CoreWeave from asset impairment risk is if the contracted revenue are linked to the specific GPU in which it seeks to invest in today. With CoreWeave planning $23 billion in capex this year, which will likely be fully deployed towards procuring Nvidia’s Blackwell systems, the corresponding contracted revenues would need to be directly linked to those chips to justify the investment.

However, if $23 billion in new long-term contracts are signed this year and realizable over six years – but transferrable to newer, more advanced GPUs – then the recoverable value of today’s Blackwell-related capex faces inevitable impairment risk. This is due to the anticipated rental rate haemorrhage once Blackwell passes its two-year peak cycle, making it increasingly difficult for revenue from the remaining four years of useful life to contribute towards recovering initial deployment costs.

This is likely the case facing CoreWeave. It’s just not economically viable for any of its customers – especially not Microsoft, its largest customer – to make such a sizable long-term commitment to one GPU series when they face imminent obsolescence within two years. Doing so would be akin to knowingly committing to an onerous contract at risk of lease asset impairment – not to mention the active breach of fiduciary duty to investors.

So, you may ask – how is ASC 842 lease accounting relevant?

CoreWeave has disclosed in its S-1 and latest Q1 2025 10Q filings that Microsoft remains its largest customer. As of March 31, 2025, Microsoft contributed to 72% of CoreWeave’s revenues. This illustrates a consistent uptrend from 35% in the year ended December 31, 2023, and 62% in 2024. Since Microsoft contributes to the majority of CoreWeave’s revenues, a quick check into its lease accounting dynamics would reveal whether CoreWeave’s success-based capex strategy is sustainable – and whether the six-year useful life attributed to its GPU assets is valid.

Remember, finance leases represent long-term lease contracts that exhibit essence of ownership assumption – either by encompassing a lease term that spans over the majority of the underlying asset’s remaining useful life or by encompassing a total lease payment value that is equal to or exceeds the fair value of the underlying asset. For instance, if Microsoft commitments to Blackwell capacity from CoreWeave for five of the chip’s designated six-year useful life, or if the committed rental payments are equal to or exceed the system’s underlying fair value today, then it’d be accounted for as a finance lease. This would be accordingly reflected on Microsoft’s balance sheet as a finance lease ROU asset typically embedded within its PP&E balance, with a matching finance lease liability typically embedded within other current and long-term liabilities.

As mentioned earlier, it’s unlikely Microsoft would want to commit to six years of Blackwell rental capacity today. However, Microsoft may be incentivized to enter into a shorter-term, high-value rental contract with CoreWeave that covers the entirety of Blackwell’s current fair value in order to offset some of its near-term supply shortfall from Nvidia. As such, it’s worth examining how much Microsoft’s finance lease ROU assets have changed in the past year to estimate the scale of GPUs it has effectively “acquired” through rentals – and how that compares to CoreWeave’s planned capex for 2025.

Based on Microsoft’s FY 2024 10K, the amount of finance lease ROU assets embedded within its PP&E balance was $25,862 million as of June 30, 2024, up 63% y/y compared to the muted 13% y/y increase observed in FY 2023. The disclosure specifies that its operating and finance leases are primarily related to “data centers, corporate offices, research and development facilities, Microsoft Experience Centers, and certain equipment”. The substantial increase in finance lease ROU assets in FY 2024 is likely inclusive of GPU rentals and data center leases, given Microsoft’s heightened focus on AI infrastructure buildout in recent years:

“Cloud and AI related spend represents nearly all of total capital expenditures. Within that, roughly half is for infrastructure needs where we continue to build and lease data centers that will support monetization over the next 15 years and beyond. The remaining cloud and AI related spend is primarily for servers, both CPUs and GPUs, to serve customers based on demand signals.” - Microsoft Q4 FY 2024 Earnings Call

This is also corroborated by management’s brief commentary during the Q4 FY 2024 earnings call on expectations of a greater mix of finance lease capex going forward to support the ongoing AI infrastructure build-out:

“The other thing, I would note, Kash, is you'll also notice there's a growing distinction between our CapEx number, and on occasion, the cash that we pay for PP&E and you're going to start to see that more often in this period because it happens when we use leases. Leases sort of show up all at once. And so you'll see a little bit more volatility.” - Microsoft Q4 FY 2024 Earnings Call

Finance lease ROU assets have continued to increase through year-to-date FY 2025. As of the nine months ended March 31, 2025, Microsoft included $37,625 million of finance lease ROU assets within its PP&E balance. That’s an increase of $11,763 million in the nine months since fiscal 2024 year-end. Hypothetically, consider half of the $11.8 billion of new finance lease ROU asset additions at Microsoft are fully attributable to server GPUs, and all of it goes to CoreWeave. That would still only amount to about 26% of CoreWeave’s planned $23 billion capex for 2025 – which deviates significantly from management’s confident confirmation that the amounts to be deployed are backed by sufficient long-term contracts to ensure breakeven.

Now, let’s consider a more aggressive scenario:

Microsoft currently anticipates about $80 billion of total capex in FY 2025 – likely on the way to finish the year at $85 billion. Based on the amount attributable to finance leases observed through Q3 FY 2025, which averages about 26% of quarterly capex, it’s likely Microsoft will add about $22 billion of finance lease ROU assets this year. By applying the same assumption to Microsoft’s FY 2026 capex, which is estimated at more than $90 billion given management’s expectations for capital spending to increase over the next year, its finance lease ROU asset additions would likely reach $24 billion in the next 12 months:

Management has also disclosed during Microsoft’s Q4 FY 2024 earnings call that half of capex is primarily attributable to servers – including both CPUs and GPUs. Half of the anticipated finance lease ROU asset addition for FY 2025 would represent $11 billion. Hypothetically, even if all of it goes towards CoreWeave GPUs, it’d still be a significant shortfall from the $23 billion in capex that the neo-cloud operator’s prepared to deploy this year. In other words, only a little under half of CoreWeave’s $23 billion GPU investments this year are secured by a breakeven contract from Microsoft at best.

Arguably, CoreWeave must’ve forged long-term lease contracts with other hyperscaler customers as well. This is supported by the consistent increase in operating lease ROU asset additions – or long-term lease contracts that don’t make up the majority useful life and/or value of the underlying asset – across hyperscalers in recent years:

While operating leases would only account for limited parts of a GPU’s useful life and value, they could potentially add up across hyperscalers to help CoreWeave breakeven on the asset. But here’s the issue – Microsoft currently accounts for 72% of CoreWeave’s revenues, with the metric consistently growing.

This means all of CoreWeave’s other customers combined currently represent less than a quarter of its revenue. And based on disclosures within CoreWeave’s latest Q1 2025 10Q, no other individual customer accounts for more than 10% of its revenue, further challenging the economic viability of its planned $23 billion capex spend this year.

As a result, it’s highly unlikely that the rest of CoreWeave’s customer base – generating just about 25% of its revenue – can deliver sufficient accretion to ensure the recoverability of $23 billion worth of 2025 GPU investments over the next six years. This would be consistent with Microsoft’s modest finance lease commitments attributable to server GPUs, implying limited incentive to lock-in long-term on third-party facilitated hardware that’s bound to become obsolete within two years. It’s further corroborated by the dominant portion of capex earmarked towards direct PP&E purchases observed across hyperscalers, highlighting their preference for full ownership of GPU assets instead to scale both external and internal workloads over the longer-term.

Although CoreWeave’s recently secured an $11.9 billion five-year contract and $4 billion four-year contract with OpenAI, there’s also no guarantee they’re tied specifically to Blackwell deployments either. The hesitation would be consistent with the rapid pace of LLM advancements observed at OpenAI, which likely coincides with preference for the latest and greatest hardware that’s currently improving at a “one-year rhythm” according to Nvidia. This accordingly weakens prospects of recoverability for CoreWeave’s $23 billion capex this year further.

The totality of facts considered would imply that as much as half of CoreWeave’s $23 billion in planned GPU investments this year are at risk of partial impairment, with the ensuing write-off estimated at $6 billion.

There just isn’t sufficient evidence drawn from hyperscaler and enterprise lease commitments that CoreWeave’s $23 billion GPU investments planned for this year can be fully recouped over a six-year useful life nor within the next two years to dodge obsolescence. Although management doesn’t typically talk about this, it’s a big enough threat to be included in the company’s latest “Business and Industry” risk disclosures within the Q1 2025 10Q:

Taken together, it’s evident CoreWeave’s six-year depreciation schedule and success-based capex strategy is not durable nor appropriate for its business model. Any subsequent impairment, which CoreWeave’s prone to, would impact its earnings outlook – something that the stock’s current valuation premium leaves no room for.

More importantly, the adverse dynamic would also subject CoreWeave to heightened liquidity risk. The company’s success-based capex strategy is primarily upheld by leverage due to CoreWeave’s ballooning losses and lack of self-sufficiency in funding operations today. Although management has painted CoreWeave’s business model as “invest first, then scale” and recover over the average asset useful life cycle of six years, that doesn’t seem to be the real-world case based on long-term lease dynamics observed across its core hyperscaler customers. Most recently, CoreWeave’s raised $2 billion in a senior notes offering at 9.25% due 2030. Yet that’s barely a drop in the bucket based on the incremental debt required to sustain its so-called “success-based” capex strategy in the coming years.

As a result, the CoreWeave stock faces significant downside risks from current levels. In the event of a correction to adjust for its potentially overstated GPU asset values, Nvidia would also be exposed due to its 7% stake in CoreWeave. It would also be detrimental to current investor confidence over the forward AI infrastructure spending trajectory – something that Nvidia’s $4 trillion market cap leaves no room for.

Conclusion

ASC 842 lease accounting has become more relevant than ever before in the tech industry. Not only because it’s embedded in hyperscaler capex forecasts – a figure that investors have been obsessed with – but also due to its increasing relevance in evolving dynamics across the AI value chain. Historically, the finance lease component of hyperscaler capex has primarily been related to data center, corporate office and other equipment rentals. But rapid AI infrastructure buildout has also exposed finance lease relevance to an emerging corner of GPU rentals. This has accordingly increased finance lease’s significance in gauging growth prospects of key picks-and-shovel players like Nvidia and CoreWeave.

But it’s likely market hasn’t quite grasped the concept, as evidenced by Wall Street estimates that take the entirety of both hyperscaler and neo-cloud capex into inferring Nvidia’s growth outlook. Although these numbers imply strong demand for AI GPUs, it doesn’t mean the fundamental underpinnings of Nvidia’s valuation aren’t overstated due to double-counting.

ASC 842 lease accounting has also revealed cracks in CoreWeave’s capex strategy and depreciation cycle. While it’s easy to be distracted by CoreWeave’s multi-fold growth, current lease dynamics observed across its key hyperscaler customers suggest the company’s business model is inherently prone to impairment. This would imply an impending accounting-related valuation reset at CoreWeave. And Nvidia wouldn’t be off the hook either, given exposure via its 7% stake in CoreWeave.

It’s evident both implications learned from lease accounting – something that most tech investors do not pay attention to – yields two material headwinds to Nvidia that its valuation today has yet to de-risk for. Be informed, and don’t be caught off guard.

Disclaimer: This analysis is for informational purposes only and represents the opinions of Livy Research. It is not investment advice nor a recommendation to buy or sell the securities discussed.

There was no better way to start finding answers to these simple (yet simultaneously complex) questions than to start digging in retrospect. What exactly had been happening at these companies in the year(s) that led up to their respective multi-bagging journeys?

My quest was to answer these questions, and use that as the roadmap for building my portfolio of fundamental-driven long ideas going forward. Specifically, I wasn’t interested in catalyst-driven upsurges, which were either typically short-lived false dawns or re-rates driven by qualitative factors that aren’t tangibly quantifiable yet. In other words, I was interested primarily in what happened in the numbers across these companies that’d led to their respective multi-fold upsurges within the span of just 12 months. Qualitative considerations were important too in building my roadmap for finding fundamental-driven long ideas, but I view them as secondary, though critically complementary, to quantitative commonalities observed.

Across the 500+ stocks listed on the NYSE and Nasdaq that had multi-bagged in either 2023, 2024 or in both of those years, my research narrowed down those within the technology sector. There were 96 technology sector stocks listed on the NYSE and Nasdaq that had multi-bagged in either 2023, 2024 or in both of those years.

While it’s impossible to pinpoint an exact “one size fits all” mathematical formula that can absolutely dictate the next set of tech multi-baggers, one thing was for sure coming out of my research work – investors’ appetite had structurally shifted from a short stint of “growth at all costs” during the non-sustainable post-COVID stimulus and 0% financing cost appreciation cycle, to a preference for “earnings or bust” set-ups that have shown relative resilience against evolving market dynamics.

Essentially, from a fundamental perspective alone, valuations have been screaming “earnings is king” in the past two years.

As a result of my research and analysis work, I’ve curated the following roadmap to guide my methodology in building a portfolio of fundamental-driven long ideas. It’ll dive specifically into what my research entails; findings and observations; the summary of fundamental trends and commonalities observed across the tech multi-baggers; and how I’ve curated a selection filter for my upcoming portfolio of fundamental-driven long ideas, which I’ll also be sharing with you soon.

My Research and Analysis Method

As mentioned in the earlier section, there were 500+ names across the NYSE and Nasdaq that multi-bagged in either 2023, 2024 or in both of those years:

Since my focus is to find tech-focused fundamental-driven long ideas, I’d filtered the population of multi-baggers observed across 2023 and 2024 to include only those within the technology sector. There were 96 technology sector tickers across the NYSE and Nasdaq that had multi-bagged in either 2023, 2024 or in both of those years:

Unsurprisingly, the technology sector was one of the major contributors to multi-baggers observed in each of 2023 and 2024. It was also the leading contributor of stocks that had multi-bagged in both of 2023 and 2024:

A high-level explanation for the technology sector’s leading contribution to multi-baggers observed in 2023 and 2024 is due to the constituents’ higher-growth nature and inherently elevated scalability that have typically yielded generous profit margins (e.g., software industry). Why these might represent leading indicators of value appreciation prospects is discussed in later sections. The observation also explains why the Nasdaq exchange delivers more multi-baggers compared to the NYSE, given constituents in the latter exchange are typically in the mature/stable stage of the business cycle and/or partake in capital-intensive industries with inherently slim profit margins.

After I’d determined the set of technology sector multi-baggers in each of 2023 and 2024, I proceeded to curate a list of quantitative and qualitative considerations that I believe to be critical for fundamental-driven long ideas. And I will be assessing said factors across each of the tech multi-baggers observed in 2023 and 2024 to determine what matters most.

Quantitative Factors – It was important that I started looking at fundamentals in the year before and after each of the stocks had multi-bagged to get a full picture of whether retrospective or forward (or both) performance was the driver of multi-fold upsurges. This meant digging back to and assessing the actual fundamental performance of the 2023-2024 technology sector multi-baggers in calendar years 2022 through 2024, as well as considering the 2025 estimates. The following are a list of four critical fundamental components I’ve assessed:

• Revenue growth: Investors typically prioritize companies that demonstrate sustainable robust growth, as this represents organic scalability from the underlying business. Organic scalability is an especially critical consideration when assessing tech companies, as the majority still remains in the start-up and growth phases of the business cycle, and only those that can scale will generate value. Scalability also reflects their ability to expand operations efficiently, which propels sustained margin expansion and cash flows growth critical for value appreciation.

• Earnings growth: In addition to scalability, operational optimization is another critical factor for start-ups and growth-stage businesses in particular. And earnings growth represents a key indicator of these two components’ progress. When earnings growth is considered in combination with revenue growth, it also reveals margin expansion or compression dynamics, offering insights into whether the company’s business model and strategy enhances or undermines the fundamental outlook that underpins valuations.

• Cash: Examining changes in cash balances provides critical insights into how well the organic operational cash burn run-rate is being managed at growing tech companies (i.e., excluding the impact of mergers and acquisitions and/or financing activities). Beyond revenue and earnings growth, changes in a company’s cash balance also reveal which part of the broader business cycle it’s in (i.e., start-up, growth, maturity or decline), and the self-sustainability and scalability of operations, which are key drivers of value appreciation. Most importantly, analyzing cash alongside capital expenditure (“capex”) is a key telling tale that answers one of the most critical questions in valuation: Is return on investment (“ROI”) expanding to support sustainable growth and efficiency?

• Capex: Capex represents a critical cash outflow that, like changes in cash balances, indicates a company’s position in its business cycle (e.g., investment vs. monetization stage). It provides deeper insights into whether revenue and earnings growth are sustainable and scalable because of investments deployed back into the organic business, or because of costly merger and acquisition (“M&A”) activities instead to artificially boost expansion. The latter scenario (i.e., M&A activities) plays a lesser role in determining the next fundamental-driven multi-bagger if it isn’t considered in concurrence with a positively quantifiable synergy. The consideration eventually leads back to whether the company’s capex strategy can help the organic business scale and optimize more efficiently to benefit cash flows underpinning valuations.

Qualitative Factors – In addition to quantitative factors, I’ve also determined a set of qualitative factors that play a critical role in dictating fundamental-driven value appreciation prospects. I categorize these qualitative factors into two groups based on their respective impact (i.e., direct vs. supplementary) on quantitative consideration factors discussed in the earlier section – namely, 1) direct qualitative considerations, and 2) supplementary qualitative considerations. The qualitative factors I consider in conjunction with the primary quantitative factors listed above include the following:

• Changes in leadership or governance: This includes a change in CEO and/or an overhaul of the board composition (e.g., addition of an activist board member). I consider these changes as direct qualitative considerations given their far-reaching impact on a company’s organic top- and bottom-line growth trajectory. Typically, changes in management aim to either reverse the sins of the past or push the company into overdrive to advance to the next phase of the broader business cycle. And either way, it’s bound to result in changes to the company’s business plan, which would inadvertently levy a direct impact on its fundamental trajectory and valuation outcome.

• New product release: Another key direct qualitative consideration is whether the company had introduced new product(s) in the year leading up to, of, or after its stock had multi-bagged. This is because new product releases typically have a direct impact on a company’s organic top- and bottom-line growth. Sometimes, the new product launch might become additive to a company’s growth outlook, given it unlocks a new total addressable market (“TAM”) or serviceable addressable market (“SAM”) for the business. In other times, there have been scenarios where a new product launch alters the fundamental direction of the company (e.g., from cost-heavy research and development to monetization-heavy go-to-market), which is typical across capital-intensive tech industries today, including electric vehicles (“EV”), autonomous mobility (“AV”) and robotics.

• Changes in business model/go-to-market (“GTM”) strategy: A change in the business model typically coincides with new management installations and/or new product releases, and levies a direct impact on a company’s top- and bottom-line growth prospects. For instance, if a company switches from a SaaS business model to a consumption-driven business model, the determinants of its key quantitative factors become significantly altered as growth and earnings considerations can be drastically different between the two strategies. A change in the GTM strategy can also yield a similar direct impact on a company’s fundamental outlook. Take Palantir Technologies Inc for example. The company’s deployment of its legacy Foundry and Gotham operating systems used to be a very company-specific sales endeavour. But since its launch of the Artificial Intelligence Platform (“AIP”) product in 2023, Palantir has primarily relied on “AIP Bootcamps” to land new and expand existing businesses in bulk and with speed. This has essentially been a significant transformation to Palantir’s GTM strategy that’s unlocked huge sales and market cost-savings and enabled substantial improvements to the scalability of its business, effectively altering the trajectory of its pace of earnings expansion in recent years.

• M&A activities – Unlike catalyst-driven trade ideas, which would typically consider potential post-M&A synergies that might not yet immediately be quantifiable, I wanted my fundamental-driven long ideas to be founded strictly based on organic company performance factors. As a result, M&A activities are considered a supplementary qualitative consideration that must be contemplated alongside the post-integration quantitative trajectory in determining their fundamental weight in creating multi-baggers.

• Changes in regulation, competitive dynamics, and the secular demand environment: These qualitative factors are usually key drivers of catalyst-driven shocks to stock valuations due to their typically speculative nature. But sometimes, they can also underpin sustained re-rates in valuation, primarily when said qualitative factors can be supplemental to a tangibly quantifiable impact on the underlying business’ growth trajectory. As a result, I consider these indirect qualitative factors when analyzing the potential key drivers of past multi-baggers in the tech sector. In these circumstances, I consider whether the related qualitative changes in consideration have been accommodative and/or additive to a company’s quantifiable fundamental trajectory. This would determine the extent of which the said qualitative factors have contributed to the multi-fold value appreciation of past multi-baggers in the tech sector, and whether they’re important factors to consider when curating my portfolio of fundamental-driven long trade ideas.

Observations

73 out of the 96 tech-sector multi-baggers (76%) observed in each of 2023 and 2024 have demonstrated a direct correlation to significant improvements in the quantitative consideration factors discussed in the earlier section. This cohort will be referred to as “fundamental-driven multi-baggers”. And 46 of them were complemented by adjacent qualitative indicators discussed in the earlier section. See Exhibit 1.

23 of the 96 tech-sector multi-baggers (24%) observed in each of 2023 and 2024 are considered “anomalies” that’ve demonstrated weak correlation to tangibly quantifiable fundamental factors, and have instead shown a strong link to speculative qualitative factors.

Quantitative Considerations

See Exhibit 2.

Recently Turned Profitable

Having recently become profitable was a key commonality across the 73 fundamental-driven multi-baggers in each of 2023 and 2024. Of the 73 fundamental-driven multi-baggers, 40 of the companies (55%) had just recently turned profitable in either 2023 or 2024, or is in process of becoming profitable in 2025.

And 31 of these had multi-bagged in the same year they’d turned profitable. 14 of which also showed significant value appreciation in the year leading up their earnings transition, and four of which continued to show significant value appreciation in the year following the switch to profitability. 13 of which demonstrated significant value appreciation only in the same year in which it turned profitable. See Exhibit 3.

Multi-Fold Annual Earnings Growth

55 of the 73 fundamental-driven multi-baggers (75%) demonstrated multi-fold annual earnings growth in either 2023 or 2024, and/or is estimated to deliver multi-fold annual earnings growth in 2025. Every company that had turned profitable recently, as discussed in the earlier section, had also demonstrated multi-fold earnings growth in the same year. 15 companies that had delivered multi-fold annual earnings growth did not demonstrate a switch from annual losses to profitability in 2023 nor 2024, as they were already profitable. This implies that profitability is a key prerequisite for multi-baggers.

43 of said 55 companies had multi-bagged in the same year they delivered multi-fold earnings growth. 15 of which also showed significant value appreciation in the year leading up to the multi-fold annual earnings growth, and four of which continued to show significant value appreciation in the year following the delivery of multi-fold annual earnings growth. 24 of which demonstrated significant value appreciation only in the same year in which it delivered multi-fold annual earnings growth. See Exhibit 4.

Significant Revenue Acceleration

18 of the 73 fundamental-driven multi-baggers (25%) showed no shift from annual losses to profitability, nor did they deliver annual multi-fold earnings growth in 2023 or 2024. They are also not expected to shift from annual losses to profitability, nor deliver annual multi-fold earnings growth in 2025.

Instead, these 18 companies demonstrated significant annual revenue acceleration in 2023, 2024, or both 2023 and 2024. Some are also expected to deliver significant annual revenue acceleration in 2025. More than half have or are expected to deliver year-on-year revenue growth of 100%+.

Significant revenue acceleration is not a must-have quantitative indicator in the 55 multi-baggers that’ve demonstrated (or is expected to demonstrate) multi-fold annual earnings growth. But significant revenue acceleration has shown itself as a key marker across multi-baggers that did not deliver multi-fold annual earnings growth. See Exhibit 5.

Qualitative Considerations

The data showed that new product launch represents a key common qualitative indicator across the 73 fundamental-driven multi-baggers observed. 40 of the 73 fundamental-driven multi-baggers (55%) had a new product launch in either 2023, 2024, both 2023 and 2024, or coming 2025. See Exhibit 6.

17 of which had launched a new product in one of 2022, 2023 or 2024, or in all of the three preceding years without correspondence with new external financing arrangements, M&A activities, management and/or governance transition, nor implementation of a new business model. Meanwhile, 23 of the fundamental-driven multi-baggers that had launched a new product during the data set period had done it alongside access to new financing arrangements, M&A activities, executive management transition, and/or the implementation of a new business model. Specifically, 8 of the 23 had coincided with new financing undertaken through a securities issuance and/or bank borrowings in one of the preceding three years, and 9 had coincided with both M&A activities and new financing acquired within the same period.

Exchange Considerations

As discussed in the earlier section, the tech- and growth-heavy Nasdaq has delivered more multi-baggers than the NYSE did in each of 2023 and 2024.

19 of the 73 assessed fundamental-driven multi-baggers (26%) were listed on the NYSE. The data shows that seven of them had shown significant value appreciation in the year leading up to the underlying business’ delivery of critical fundamental improvements (e.g., significant revenue acceleration; transition to profitability; multi-fold annual earnings growth). This is likely due to the fact that companies listed on the NYSE are typically viewed by investors as more mature and stable, which is corroborated by the fact that the exchange has consistently outperformed broader markets during times of elevated volatility. The observation accordingly increases confidence and visibility into the reliability of forward fundamental projections, hence potentially explaining why NYSE-listed fundamental-driven tech multi-baggers have typically appreciated by multi-fold in the year leading up to their underlying business’ delivery of critical fundamental improvements versus in the year of.

There were another seven NYSE multi-baggers that had shown significant value appreciation in the same year that their underlying businesses had delivered critical fundamental improvements. Three of the NYSE-listed fundamental-driven tech multi-baggers started delivering significant value appreciation in both the year before and the year of their respective critical fundamental improvements, which aligns with the earlier discussion that there tends to be greater investor confidence in NYSE estimates. Two of the NYSE-listed fundamental-driven tech multi-baggers extended their significant value appreciation in the year after they’d delivered the critical fundamental improvements. See Exhibit 7.

Meanwhile, 54 of the 73 assessed fundamental-driven multi-baggers (74%) were listed on Nasdaq. Unlike the fundamental-driven tech multi-baggers observed in the NYSE, those listed on the Nasdaq had primarily (23 of them) shown significant value appreciation in the same year they’d delivered critical fundamental improvements. This is likely because Nasdaq-listed stocks have historically been more volatile than broader market performance, leading investors to prefer action-taking when there’s greater evidence supportive of value appreciation opportunities (e.g., responsive valuation upsurges after periodic earnings outperformance on Nasdaq vs. pre-emptive valuation upsurges prior to periodic earnings outperformance on the NYSE).

15 showed significant value appreciation in the year before and in the same year they’d delivered their critical fundamental improvements. Only 12 showed significant value appreciation only in the year leading up to their delivery of critical fundamental improvements. Three showed significant value appreciation in the same year and the year after they’d delivered their critical fundamental improvements. And one showed significant value appreciation only in the year after they’d deliver critical fundamental improvements. See Exhibit 8.

Anomalies

There have been 23 tech multi-baggers observed in each of 2023 and 2024, or in both 2023 and 2024 that’ve shown limited correlation to the critical fundamental improvements (e.g., shift from annual losses to profitability; multi-fold annual earnings growth; significant revenue acceleration). See Exhibit 9.

21 of which (91%) have demonstrated significant correlation and sensitivity to one of the following key market themes since 2023: artificial intelligence, quantum, crypto, autonomous mobility, and China.

• Artificial Intelligence (“AI”): Nine of the anomalies have demonstrated direct participation in AI opportunities, whether it’s directly through AI product launches (e.g., infrastructure hardware; foundation large language models; end-user AI applications) or indirectly through the supply of data and data management solutions.

• Quantum: Five of the anomalies have demonstrated direct participation in the quantum rally that began in late 2024. This coincided with Alphabet Inc’s introduction of its first “Willow” quantum chip, which demonstrated capabilities in solving complex problems that should’ve taken longer than the universe’s lifetime in under five minutes. Yet this quantum rally was short-lived and mostly fizzled in early 2025 when NVIDIA Corp CEO Jensen Huang estimated that quantum computing will not have useful mass market use cases for at least another two decades. This announcement pretty much solidified concerns that the upsurge lacked durability given limited quantifiable fundamental support. The five quantum-related multi-bagging anomalies in 2024 have lost at least -43% of their respective values in the first two months of 2025 (average -60% YTD declines as of February 28, 2025).

• Crypto: Three of the anomalies have direct participation in the crypto industry. Crypto names have demonstrated significant volatility in recent years, given their inherently elevated sensitivity to macroeconomic conditions amidst the monetary tightening cycle in recent years, as well as changes to the regulatory environment (e.g., in anticipation of BTC spot ETF approvals in January 2024; impending regulatory clarity for crypto assets under new Trump administration in late 2024). Yet the majority of these stocks have demonstrated limited correlation to tangible fundamental improvements, given nominal monetizable use cases in the mass market for cryptocurrency. Two of the three crypto-related multi-baggers observed in each of 2023 and 2024 have declined by more than 30% from their respective year-end peaks within one week. Only Strategy (formerly MicroStrategy Inc) has maintained its multi-fold gains since 2023, through it’s primarily due to Bitcoin accumulation funded through external financing rather than organic fundamental business growth.

• Autonomous mobility: Two of the anomalies have direct participation in the reemergence of the autonomous mobility theme following the new Trump administration’s promise to prioritize the related regulatory framework for public rollouts. However, since the anticipated improvements to regulatory clarity remain uncertain, there’s still limited visibility into the timing and size of which autonomous mobility opportunities will materialize. This has accordingly resulted in weaknesses to the durability of related upsurges, which is consistent with the approximate average decline of -34% over the first two months of 2025 (i.e., since the related multi-bagger anomalies’ respective peaks in late 2024 to date).

• China: Two of the anomalies are China American depository receipts (“ADRs”) listed on Nasdaq, which demonstrated sudden, yet non-sustainable, upsurges that coincided with favourable changes in central government policies. One of the names that’d multi-bagged in early 2023 surged amidst China’s post-pandemic reopening rally, while its US counterparts struggled with one of the most aggressive monetary policy tightening cycles in four decades. The other name that’d multi-bagged in late 2024 surged in response to Beijing’s improving policy support for the local economy at the time, as well as the announcement of a planned acquisition; it’s also been a key beneficiary of the emerging Chinese AI rally unleashed by DeepSeek’s breakthrough efficiency revelations. Yet both anomalies have demonstrated limited improvements to critical fundamental considerations discussed in the earlier sections.

The sharp positive response to favourable policy changes, although not yet fundamentally quantifiable, is likely reflective of Chinese ADRs’ inherently significant discount on a relative basis to comparable American counterparts due to the China risk premium driven by regulatory, geopolitical and macroeconomic uncertainties. This accordingly leaves a generous margin for multiple expansion in response to qualitative factors that may be supportive of reducing the China risk premium. However, related value appreciation have typically shown lacking durability due to limited evidence of tangible fundamental improvements.

Pintec Technology Holdings Limited, which multi-bagged in 2023, acquired most of its gains in the first four months of the year, which coincided with China’s post-pandemic reopening rally. It then pared the gains by as much as 95% within just one month from the YTD peak reached in April 2023, and ended the full year with a 172%+ appreciation instead. Meanwhile, iClick Interactive Asia Group Limited, which multi-bagged in 2024, acquired most of its gains in the last two months of the year, which coincided with Beijing’s slew of economic support announced since, as well as its planned acquisition of digital wealth management services firm Amber DWM Holding Limited. Neither stocks have demonstrated significant organic fundamental improvements in the underlying business to match their qualitative-driven momentum.

There are two other anomalies that have demonstrated key qualitative considerations – namely, a new CEO installation (i.e., leadership/governance change) and a new product launch – in the year corresponding to their respective multi-bagging journeys. Although they did not coincide with critical fundamental improvements (e.g., a shift from annual losses to profitability; multi-fold annual earnings growth; significant revenue acceleration), both companies have shown consistent earnings expansion from 2022 through 2024, which is indicative of scalability and optimization – two critical business components as discussed in the earlier sections.

Based on my assessment of the anomalies and potential drivers of their multi-fold value appreciations, the majority of which have been sudden and short-lived shocks. As a result, these are akin to catalyst-driven trades, in my opinion, which are typically reflective of re-rates driven by qualitative factors that are not immediately quantifiable yet, leaving little fundamental support required for a sustained valuation appreciation. I aim to explore further on drivers of catalytic trade ideas in a future analysis.

Control Group Analysis

I’ve also assessed names that exhibited critical fundamental improvements (e.g., multi-fold revenue acceleration; multi-fold earnings growth; transition from annual losses to profitability) in 2023 and/or 2024, and/or are expected to deliver said characteristics in 2025, yet did not multi-bag in 2023 nor 2024. This procedure was performed to determine any other quantitative and/or qualitative weaknesses present that could’ve offset the positive impact of critical fundamental improvements on valuations.

The key takeaway was that GAAP-based profitability was a key prerequisite for significant value appreciation. This compares to an earnings transition marked by publicly available EPS data, which typically includes non-GAAP adjustments. Another key consideration is that the presence of critical fundamental improvements (e.g., multi-fold revenue acceleration; multi-fold earnings growth; earnings transition) must be sustainable. In other words, the critical fundamental improvements must not reverse in the following period (e.g., revert to annual losses and/or significant revenue deceleration) or be driven by one-time events (e.g., M&A activities); any signs of slashed guidance from management would also represent a downside risk to value appreciation prospects. This accordingly reflects the mission-critical role of organic scalability and optimization in the underlying business for sustainable fundamental-driven value appreciation.

The below section will provide further details on observations in the control group analysis:

Nasdaq

Multi-Fold Revenue Appreciation

In the Nasdaq exchange, there were four names that delivered multi-fold revenue growth in 2023, and one name that is expected to deliver multi-fold revenue growth in 2025. Yet none of the five names have delivered significant value appreciation in 2023 nor 2024.

Two of the names demonstrated multi-fold revenue appreciation in 2023 due to full-year integration of newly acquired companies. As discussed in the earlier section, M&A-related changes in fundamental trajectory aren’t considered characteristics of organic scalability and/or optimization. This likely explains the absence of accompanying multi-bag value appreciation at the two names.

Two of the other names that demonstrated multi-fold revenue growth in 2023 remain unprofitable. Based on my assessment of 2023 and 2024 tech multi-baggers in the earlier section, it appears that profitability is a key prerequisite for significant value appreciation. This is consistent with expectations that tangible cash flow accretion at a profitable company is a key driver of sustained value appreciation.

The fifth name is expected to deliver significant revenue growth in 2025 (Nebius Group NV). As discussed in the “Exchange Considerations” section, valuation changes in Nasdaq-listed names are typically responsive rather than pre-emptive to earnings results. The company is also not yet profitable, nor is it expected to turn profitable in 2025. Hence this could explain why the stock has yet to deliver significant value appreciation.

Multi-Fold Earnings Growth and/or Earnings Transition

There were 47 tech names on Nasdaq that exhibited multi-fold earnings growth and/or transitioned from annual losses to profitability in 2023, yet were not accompanied by significant value appreciation in 2023 nor 2024. 23 of these names have or are expected to backtrack into earnings declines or losses post-2023, which contradicts the requirement for organic scalability and optimization to underpin sustained value appreciation. And 15 of these names delivered net ~25% value appreciation in 2023 and 2024. 13 of which had delivered more than 50% appreciation in 2023, which is the same year they delivered multi-fold earnings growth, before paring gains in 2024. Although they did not multi-bag, the net 25%+ appreciation through 2023 to 2024 is a threshold I consider as reflective of the reward for their partial fundamental improvements.

This leaves nine anomalies to be further assessed. A deeper dive shows that five of them weren’t profitable on a GAAP basis (recall that the data assessed includes non-GAAP EPS metrics), which contradicts a key prerequisite for multi-fold value appreciation. Three of the nine anomalies also slashed key earnings and growth expectations for full-year 2024, which likely impacted their value appreciation prospects as well. One of the anomalies was a Chinese online recruitment business, which suffered significantly from China macroeconomic weakness and consistent underperformance in estimates; this accordingly increased uncertainties to the scalability of the underlying business. And one of the anomalies (First Solar Inc) had actually multi-bagged in 2022 in anticipation of its earnings transition achieved in 2023.

Meanwhile, 28 names listed on Nasdaq exhibited multi-fold earnings growth in 2024, but did not deliver significant value appreciation in 2023 nor 2024. 24 of the names delivered net ~25% growth in 2023 and 2024; 16 of which delivered 50%+ value appreciation in 2024. Three of the names are expected to return to earnings declines in 2025, indicating the lack of scalability in the underlying business, which is another key prerequisite for value appreciation.

This leaves one anomaly to be further assessed – namely, Ultra Clean Holdings Inc (“UCTT”), which is a critical solutions provider to the semiconductor industry. The company delivered multi-fold EPS growth in 2024 and is GAAP profitable. Yet the stock’s modest ~20% value appreciation observed in 2024 is likely a reflection of market’s concerns about the company’s exposure to tightening semiconductor export curbs – especially at its China business. The company generated 73% of its sales outside of the US in 2024, with management citing China as a primary revenue contributor. This accordingly raises uncertainties about UCTT’s revenue outlook, and inadvertently introduces downside risks to the durability of its earnings critical to supporting the stock’s valuation.

NYSE

Multi-Fold Revenue Appreciation

Recall that NYSE multi-baggers typically exhibit significant value appreciation in the year leading up to their respective critical fundamental improvement(s). Yet there was one name (GCT Semiconductor Holding Inc / “GCTS”) listed on the NYSE that did not deliver significant value appreciation in 2024, despite expectations for multi-fold earnings growth in 2025. And similar to observations in non-multi-baggers on the Nasdaq discussed in the earlier section, GCTS’ underperformance in 2024 is likely the result of its ongoing GAAP losses, despite expectations for impending multi-fold earnings growth.

A deeper dive into the company’s 2024 fundamentals also reveals that it’d benefitted from a de-SPAC (i.e., catalyst-driven) upsurge earlier in the year. This was coupled with a transition to GAAP profitability in 1Q24. However, the earnings transition was not sustainable, as it was the result of a one-time $14.6 million benefit to the bottom-line resulting from gains realized on the extinguishment of a liability tied to its development agreement with Samsung.

Multi-Fold Earnings Growth and/or Earnings Transition

23 NYSE-listed tech stocks that exhibited multi-fold earnings growth in 2023 did not deliver significant appreciation in 2022 nor 2023. Six of them were expecting earnings deterioration in 2024, with one other also expected to revert back to annual losses. One of them was also not GAAP profitable. These are indicators of lacking organic scalability in the underlying business, which likely explains the affected stocks’ limited value appreciation observed through 2022, 2023 and 2024.

Meanwhile, 18 of the said 23 names had either demonstrated more than net ~25% value appreciation between 2022 and 2023 or between 2023 and 2024; eight of which delivered more than 50% annual value appreciation in either 2023 or 2024. The outperformance against broader markets implies that they were still rewarded for their partial earnings improvements, despite not having multi-folded in value.

This leaves five anomalies to be further assessed. Three of which either did or were close to multi-bagging in 2023 in anticipation of their respective earnings transitions before paring gains due to slashed guidance. And two of the names faced significant deceleration from 2022 alongside repeated earnings underperformance, which likely led to a downward adjustment in valuation considerations given concerns about the durability of their respective earnings transitions.

Meanwhile, 57 NYSE-listed tech names exhibited multi-fold earnings growth in 2024, yet did not multi-bag in valuation in 2023 nor 2024. However, most (30) had demonstrated strong value appreciation in the preceding year, resulting in net gains of more than 25% in the two years through 2024; 17 of which delivered value appreciation of more than 50% in either 2023 or 2024.

In the remaining 27 names, three of which are expected to return to annual losses in 2025, and six face impending earnings declines. Five others were explained as part of the 2023 anomalies in the earlier section.

This leaves 13 anomalies to be further assessed. 10 of which were not actually GAAP profitable, despite positive EPS data assessed that likely included non-GAAP adjustments.

For the remaining three names, one of them (Teradata Corp) did not actually deliver multi-fold earnings growth on a GAAP basis, with the underlying business also facing multiple downward revisions to both 2024 and 2025 estimates alongside recurring revenue declines, thus raising further weakness in earnings and value accretion prospects.

Another name (HUYA Inc – ADR / “HUYA”) that’d delivered multi-fold earnings growth and a transition to annual profitability also had its bottom-line improvements overshadowed by persistent revenue declines. Specifically, HUYA’s multi-fold earnings growth was primarily realized from significant cost-savings resulting from a pivot in its business model to prioritize higher-margin revenue streams. However, those growth avenues face impending secular declines, which are worsened by HUYA’s significant revenue exposure to China’s mixed macroeconomic backdrop, harbingering inevitable earnings deterioration risks.

The final anomaly (Tuya Inc / “TUYA”) that’d delivered multi-fold earnings growth in 2024 but did not multi-bag in 2023 nor 2024 has actually delivered multi-fold value appreciation in early 2025. The delayed appreciation was likely due to the added tailwind of China’s improving policy backdrop in recent months, which alleviated some of the region-specific risk premium that’s been weighing on the Chinese ADR’s valuation. This has also likely bolstered the durability of TUYA’s underlying business performance and reinforced market confidence in its fundamental-driven upside potential. The company will remain on my watchlist for further fundamental-driven value appreciation prospects in the current year.

Summary of Key Commonalities that Underpin Tech Multi-Baggers in 2023 and 2024

The following ranks key commonalities observed across the 73 fundamental-driven tech multi-baggers by importance:

1. Earnings is king – The shift from annual losses to profitability, as well as multi-fold annual earnings growth is the number one common driver observed across the 73 fundamental-driven tech multi-baggers in the assessed data set. This is likely because it’s an indicator of sustained margin expansion and profitable growth, as product/services sales continue to scale. It’s essentially a marker of change in the broader business cycle, from start-up investment phase to growth monetization phase. And this accumulation of underlying free cash flows driven by sustained improvements in operating leverage represent a critical driver of future value appreciation prospects.

2. Significant revenue acceleration – The 18 fundamental-driven tech multi-baggers observed in each of 2023 and 2024, or in both 2023 and 2024 that did not deliver a shift from annual losses to profitability nor multi-fold annual earnings growth had instead demonstrated significant annual revenue acceleration. More than half had exhibited or are expected to exhibit multi-fold annual revenue growth, with the remaining six demonstrating growth acceleration at an annual rate of at least +11% y/y (average annual rate +20% y/y). This observation is also consistent with the importance of scalability and the characteristic’s impact on profit margins critical for value accretion.

3. New product launch – This has emerged as a leading qualitative factor that’s directly supplementary to critical quantitative factor improvements, such as significant revenue acceleration and/or multi-fold annual earnings growth. As discussed in an earlier section, new product launches that are additive to growth and/or enabling TAM expansion can underpin significant valuation re-rates, which can sometimes result in multi-baggers like those within the assessed data set.

Filter and Selection Process

I aim to leverage the findings from my screener of 2023/2024 tech multi-baggers discussed in the foregoing analysis to underpin the construction of my fundamental-driven long trade ideas portfolio this year. This portfolio will consist of a maximum of 10 names.

Initial Fundamental Shortlist

I will first curate a new data set that consists of the critical fundamental factors across the population of technology sector names listed on the NYSE and Nasdaq. These critical fundamental factors include revenue growth, earnings growth, period-end cash balance and capex for 2025 and 2026. The 2025-2026 assessment timeline is consistent with my goal of finding fundamental-driven long trade ideas with optimal upside potential in the current year, which, based on findings in my foregoing analysis, is typically dependent on fundamental estimates in the next 12 to 24 months (i.e., current and following calendar year).

Consistent with my findings listed in the “Summary of Key Commonalities…” section, tech sector constituents that demonstrate prospects of multi-fold earnings growth while already GAAP profitable, first-year profit recognition, and/or multi-fold revenue growth acceleration in 2025 and/or 2026 will be considered.

This shortlist will be further refined based on observations discussed in the “Exchange Considerations” section. Specifically, Nasdaq-listed candidates that are expected to deliver critical fundamental factors in the current year, and NYSE-listed candidates that are expected to deliver critical fundamental factors in the following year will be included in the initial shortlist.

Qualitative Considerations

The next step is to further refine the initial shortlist by performing a preliminary qualitative analysis on each of the selected names. This is done by screening the latest periodic filings (annual and/or quarterly reports) and the most recent earnings transcript for the complementary qualitative markers assessed in my retrospective analysis of fundamental tech multi-baggers in 2023/2024. These include announcements of a recent/upcoming product launch, new CEO installation and/or leadership and governance changes, and/or the adoption of a refined go-to-market strategy, which are complementary qualitative factors that could be indicative of additive growth and/or margin expansion.

The initially shortlisted names that also demonstrate one or more of the said complementary qualitative factors will be further prioritized. And names that are expecting an impending fundamental uplift (e.g., multi-fold earnings and/or earnings growth) within the next 12 to 24 months due to inorganic measures (e.g., M&A activities) will be removed from the shortlist.

I will also be on the lookout for downward adjustments to management guidance exhibited at names that are expected to deliver critical quantitative fundamental improvements. This is consistent with observations in the “Control Group Analysis” section, which identifies those with consistent earnings underperformance as multi-bag misses – even if they’re expected to deliver multi-fold revenue acceleration, multi-fold earnings growth, and/or a transition from annual losses to profitability.

Fundamental Deep Dive

The shortlist refined by quantitative and qualitative considerations will be subject to a fundamental deep dive, which is a two-step process that will first determine the durability of financial estimates and projections, then assess the estimated intrinsic value of the company.

i. Determine the durability of financial estimates – This primarily entails learning the shortlisted company’s business in detail to confirm the durability of fundamental estimates over the next 12 to 24 months that were assessed earlier on. This step should refine my fundamental long tech shortlist to no more than 20 names.

The purpose is to determine whether the projected earnings growth, transition to annual profitability, and/or significant revenue acceleration can be sustainably achieved through organic optimization and scalability of product/services sales. Any indications of one-time drivers (e.g., one-time contract recognitions, pulled-forward demand) would suggest that the critical fundamental consideration factor is non-sustainable and unlikely to translate into significant value appreciation; these names will be eliminated from my shortlist.

This step will also help with determining the company’s TAM, where exactly it’s at in capitalizing on related opportunities, and whether it has the means to optimize participation in the said growth opportunity. Is the company shifting from start-up (i.e., investing cycle) to growth (i.e., monetization cycle) phase of the broader business cycle at a time where significant opportunities persist in the industry? Does the company have a unique competitive advantage that can help it better navigate through the rapidly evolving business dynamics (e.g., cyclical headwinds, competition)?

The last thing you’d want is to invest into a company that’s readying to ship a new, yet inferior, product into an industry that’s already concentrated with much more superior competitors. It’s essentially a recipe for ROI Armageddon, since intensive investment outlays deployed are now met with limited monetization prospects. A prime example of which would be Intel Corp, which shipped its “Sapphire Rapids” Xeon server CPUs late into a cyclical slowdown in 2023, causing it to lose significant market share to rival Advanced Micro Devices Inc; it’d go on to repeat the same mistake by shipping an inferior Gaudi 3 accelerator in 2024, which pretty much cost Intel’s entire participation in the ongoing AI/data center boom.

This refinement process also means that the names with the highest percentage of estimated earnings increase and/or revenue increase in the NTM doesn’t automatically translate into increased confidence in its multi-bag prospect by default.

For example, consider company A, which is expected to increase earnings by 150% in 2025, and company B at 125%. They both meet the shortlist criteria of multi-fold earnings growth. Yet company A’s higher estimated rate of earnings growth doesn’t automatically mean it has a better chance of multi-bagging than company B. While company A’s higher level of earnings increase does give it a higher 50-percentage point margin for adjustments to stay compliant with the multi-fold earnings growth criterion, its prospects of becoming a multi-bagger could also depend on the estimate’s source (organic vs. M&A) and durability (scale and optimization vs. one-time cost benefits). For instance, if half of company A’s estimated earnings growth in the NTM is due to a recent acquisition, then company B would be the better pick given its value appreciation prospects are based on organic fundamental improvements. The only way to find out which is indeed the stronger candidate in becoming a multi-bagger is through the fundamental deep dive refinement process.

ii. Financial model the estimated intrinsic value (price determination) – After the financial elements check out (e.g., the critical fundamental considerations are sustained by organic scalability/optimization; significant industry growth runway; unique competitive advantages present), the final step would be to determine their estimated intrinsic value (i.e., price target). This will be key in guiding the entry and exit points for the selected fundamental-driven long trade idea.

I prefer to use the discounted cash flow (“DCF”) method, as all shortlisted names should be either approaching or are already profitable and cash flow positive. The DCF method would yield a valuation that is better reflective of the intrinsic value of underlying business cash flows prospects. Key assumptions considered in performing the DCF analysis include revenue and earnings growth estimates, which flow directly into the assessed cash flow projections; cost of capital considerations; and the terminal value.

The cash flow estimates will be determined primarily based on historical trends in revenue growth and earnings margins, as well as forward adjustments considering factors including, but not limited to, management guidance, new product launches, and changes in demand dynamics.

Cost of capital assumptions will primarily relate to the company’s capital structure (debt vs. equity) and risk profile (e.g., cost of debt vs. risk-free rate).

And the terminal value will be computed based on the application of an estimated perpetual growth rate on terminal-year cash flows. Terminal-year cash flows is proxied by projected EBITDA in the final year of the forecast period when the underlying business’ growth is expected to have reached steady-state. The estimated perpetual growth rate will mirror the projected pace of economic expansion across the company’s core operating region(s), which aligns with the rate of which steady-state cash flows can grow sustainably into perpetuity.

In some cases, the estimated intrinsic value computed under the DCF analysis can be supplemented by an estimated relative-basis value computed under the multiple-based approach. This combined valuation method can sometimes help to better gauge the shortlisted stock’s upside potential.

The multiple-based valuation approach would entail mapping stocks across a comparable industry (e.g., software) based on their earnings/revenue growth estimates against their respective P/E and/or P/S multiples. Each stock’s placement on the chart would be compared to the average regression trend line, as well as multiples exhibited by peers with a comparable growth prospect to determine its premium/discount on a relative basis. Below is an example of software industry price-to-sales multiples mapped against their respective NTM revenue growth based on Wall Street estimates as of calendar year-end 2024:

I’d utilize this combined valuation approach that considers both the DCF and multiple-based outcomes in scenarios where risks of confirmation bias are elevated (e.g., AI names have better growth acceleration prospects) and/or underlying cash flow projections are not yet reflective of the impending re-rate (e.g., a company that’s approaching profitability is not yet trading at the extended multiple of a comparable peer that already exhibits a positive bottom-line).

The top 10 names that demonstrate the highest level of durability in impending critical fundamental considerations (e.g., multi-fold earnings outlook; transition from annual losses to profit; multi-fold revenue acceleration) and complementary qualitative environment (e.g., new product launch), alongside the best price-to-value set-up will be added to my basket of fundamental-driven long trade ideas.

Names that are constituents within the Nasdaq 100 and/or S&P 500 will also be preferred. This is because the Nasdaq 100 and S&P 500 are two of the biggest indices in the world, which inadvertently garners an elevated level of global and institutional interest. As such, the incremental market demand is expected to add reinforcement to the relevant shortlisted names’ prospects of significant value appreciation. I leave this as a supplemental consideration at the end, rather than in the earlier filtration process, given Nasdaq 100 and S&P 500 constituents only had a minor representation across the population of tech multi-baggers observed in 2023 and 2024.

Conclusion

I aim to use the findings and takeaways from my retrospective dive into 2023 and 2024 tech multi-baggers to curate a fundamental-driven long portfolio that’ll be adaptive to and resilient against the evolving dynamics of forward market conditions. Based on the findings, stocks that demonstrate prospects of delivering critical fundamental considerations, such as multi-fold earnings growth and/or significant revenue acceleration, exhibit qualities of adaptiveness to evolving market conditions given their robust financial foundations. Meanwhile, those that are supplemented with additional qualitative traits, such as new product launches, could demonstrate incremental resilience against downside risks given additive growth prospects. Taken together, I will prioritize investigation into tech names across Nasdaq and the NYSE that are already profitable or are expected to turn profitable, with prospects of multi-fold earnings and/or revenue growth, over the next 12 to 24 months as the initial gateway to determining the best-performing fundamental-driven long trade ideas this year.

This analysis marks the first of two parts of my fundamental-driven long tech portfolio construction process. I’d value any feedback as I work on part II — finding the 10 constituents of my fundamental-driven long tech portfolio — which I’ll be sharing with you in a week’s time.

Thank you for reading.

Exhibits

Please click here to access the accompanying exhibits.

Disclaimer: This analysis is for informational purposes only and represents the opinions of Livy Research. It is not investment advice nor a recommendation to buy or sell the securities discussed.

Not too long ago, the Apple bulls were still chiming of an AI-driven upgrade super cycle. But that narrative’s taken a complete 360 in the last few weeks. It’s evident in the slew of recent analyst downgrades, which have likely been triggered by bleak iPhone sales data from third-party research firms – particularly in China, one of Apple’s most important markets.

However, this shouldn’t have come at such a surprise. Apple Intelligence’s lacking muscles in reaccelerating growth for Apple was blatantly evident in the details.

Between the introduction of Apple Intelligence in June and when the company reported its F4Q24 earnings results in late October, the stock had already surged 17%. Yet there was nothing to show for it. iPhone 16 upgrades were mediocre at best during the September quarter compared to the AI hype leading up to its go-to-market – and to give it the benefit of the doubt, maybe it was just because the full slate of Apple Intelligence features weren’t readily available yet at the time, and the period had just captured a week-ish of new iPhone 16 sales.

But the devil was in the details all along. Management had never guided an outsized upgrade cycle for the iPhone 16. Despite calling Apple Intelligence a “compelling upgrade reason”, management went on to do otherwise with a December quarter guidance that followed in whimper. As usual, Apple didn’t provide a direct guidance for iPhone sales. But given management’s expectations for low- to mid-single digits y/y growth in total company revenue for the December quarter, which would consist of services revenue growth in the mid-teens percentage range, it was safe to assume that iPhone sales growth for the period would be in line with historical seasonality observations – no more, and no less, not even with Apple Intelligence.

There was also a unanimous tone of doubt coming from the US big 3 telcos’ CEOs on anticipated AI-driven iPhone upgrades at home for the holiday season. Worst off, Apple was in lack of an AI strategy for China as well, while local rivals were capitalizing on the AI-driven tailwind with new innovations and homegrown operating systems that would detract users further from iOS. By then, it should’ve been evident that cracks were showing in the AI upgrade super cycle thesis that’s been buoying the Apple stock’s ballooning premium.

Yet Apple bulls were still holding steadfast on their conviction that an upgrade super cycle driven by the AI sensation was imminent – their confidence basically went something like this: if not now, then later 2025 with the iPhone 17 and Apple Intelligence for China; hell or high water, Apple’s going to win the AI race. Paired with Apple’s ongoing share buybacks, analysts were still eager in upgrading the stock’s outlook.

But by December, it was evident that entrenched market share loss in China had become Apple’s base case. It didn’t help that third-party research data was also estimating a 5% y/y decline in global iPhone sales during the December quarter – Apple’s largest sales period of the year. And at a time when 2024 global smartphone shipments had recovered from its bottom with 4% y/y growth, Apple iPhone shipments were predicted to have fallen by 2% y/y, leading it to relinquish a whole percentage point of its global market share to 18%.

Looking ahead, Apple’s positioned for nothing but an imminent one-two punch that will take the stock correction-bound.

One – based on the current situation of entrenched market share loss in China and limited AI-driven iPhone upgrades at home in the US, Apple’s likely to underperform even its own conservative growth guidance for the device. This would set a weak base for an even weaker March quarter outlook given inevitable seasonality-driven headwinds. The ensuing downward shift in Apple’s fundamental trajectory should lead to an imminent reduction of the stock’s currently lofty premium.

And two – the next wave of correction will come when Apple Intelligence goes to China. It may sound ironic, since most still assume that the software’s China debut would lead to a release of pent-up iPhone upgrades. But instead, it’s more likely that Apple Intelligence’s eventual China debut will yield no incremental iPhone demand to show for it. That’d be the moment of truth to put all of market’s remaining confidence in Apple’s AI upgrade super cycle to rest. And the downside risks won’t stop there. With the iPhone being the anchor to Apple’s broader ecosystem (particularly the higher growth and more profitable services revenue stream), a weakening installed base would also deteriorate Apple’s longer-term earnings quality and FCF outlook.

Apple’s outlook is bound to worsen until it proves a new innovation cycle’s coming along that can drive additive growth for the company. But with no immediate respite on that end, its weakening fundamentals will eventually take precedence over the superficial upgrade cycle narrative that’s been driving the stock’s valuation, leading to a fast-approaching eventual correction.

Entrenched China Market Share Loss is Now the Base Case

I switched gears from seeing entrenched China market share loss for Apple as merely a “risk” to an imminent base case “reality” in late 2024 when iPhone shipment trends in the region were consistently diverging from those observed at its local rivals. The first warning was when third-party research data showed iPhone sales during China’s hottest 11.11 shopping festival in November had fallen in the double-digits percentage range. This was followed by confirmation that foreign smartphone sales in China, a category dominated by the iPhone, had fallen by 47% y/y in November.

Blighted by intensifying competition and a lacking AI strategy to appease the crowds, Apple turned to promotions and discounts (a rare sighting) for its Chinese buyers. And while that’s done little in lifting local interest for the iPhone, the stock continued to fail in reflecting Apple’s bleak reality, considering its resilient upsurge heading towards the end of 2024.

But going into 2025 was a different reality. Industry data and analyst reports were starting to fight for airtime in proclaiming Apple’s China demise. iPhone shipments were reported to have fallen by 17% y/y in 2024, while smartphone sales from local rivals like Huawei had grown by as much as 37% y/y. Apple’s disappointment’s even more evident when compared to the 4% y/y recovery in China smartphone shipments following a two-year slump. December quarter sales – which represent the first full quarter of iPhone 16 shipments – fell by more than 18% y/y in China. The persistent weakness has effectively dropped the iPhone’s China market share from 24% in 2023 to 17% by the end of 2024.

And the outlook remains grim, with local policy support providing further favours to Apple’s Chinese rivals. Beijing was clear that boosting consumption was a priority in 2025 – but the caveat was that foreigners like Apple would be largely excluded. This was evident in the extension of consumption vouchers to purchases of smartphones, tablets and smartwatches. Eligible purchases priced at under 6,000 RMB can get a 15% subsidy, up to a maximum of 500 RMB. And for the iPhone 16, only the standard 128GB model would meet the pricing threshold, yet it’s not even Apple’s best-seller in China. It’s evident the new subsidies will only further market share gains for Apple’s local rivals at the continued expense of China iPhone demand.

Apple Intelligence for China’s also becoming a “too little, too late” strategy for the company in reviving local demand for the iPhone. The eventual rollout of Apple’s newest AI features is, ironically, likely to become the telling tale of the iPhone’s deteriorating growth outlook. It’s already evident that Apple Intelligence’s been doing little in encouraging upgrades in the US. These observations have been unanimously shared by executives across AT&T, Verizon and T-Mobile (collectively, the big 3 US telcos). And the same trend will emerge when Apple Intelligence goes to China (later this year, if its lucky in getting the regulatory greenlight from Beijing). The new features are unlikely to resuscitate local iPhone demand, especially given the uprise of rivalling AI smartphones and homegrown operating systems that continue to upend the appeal of Apple’s broader ecosystem.

And neither are the recently speculated new form factors for the next-generation iPhone 17 going to do the trick. Industry reports have been anticipating foldable and thinner “Air” variants of the iPhone 17 to debut later this year. But they aren’t going to be the answer in reviving iPhone demand in China. Since the thinner “Air” variant relies solely on eSIM technology, the design wouldn’t abide by China’s requirements for physical SIM cards with real name registration. Meanwhile, foldable smartphones are also losing their lustre amongst Chinese consumers. While Huawei has maintained its leadership in the segment, which was recently reinforced by its launch of the world’s first tri-fold smartphone, broader foldable smartphone shipments have been in steep deceleration over the past two years.

The implications of continued iPhone market share haemorrhage in China are significant for Apple, as it challenges the AI upgrade super cycle currently priced into the stock’s premium. The increasingly entrenched market share loss suffered in the region underscores vulnerability to Apple’s broader growth outlook. Given the iPhone represents the anchor of Apple’s broader ecosystem, the product’s diminishing installed base risks spilling over into adjacent components of business. This includes impacting the higher growth and more profitable services revenue stream that have been a key medium in underpinning the company’s cash flows in recent years.

Imminent Earnings Underperformance

Taken together, I anticipate Apple to underperform December quarter results at its upcoming earnings update on January 30. Remember, Apple management had never guided an outsized holiday season outlook despite the launch of AI-enabled iPhone 16s. And neither have US telco executives been preparing for significant AI-driven upgrades, nor were they shy in cautioning about the doubtful reality. The AI upgrade super cycle has always been a bull-constructed concept.

Despite management calling Apple Intelligence a “compelling upgrade reason”, the December quarter guidance was much more modest and suggested iPhone growth in the mid- to high-single digit percentage range, in line with historical observations. It also aligned with commentary from leadership across Apple’s key US carriers:

AT&T Inc. 3Q24 Earnings Call:

As expected, service revenue growth was partially offset by lower equipment revenues with a postpaid upgrade rate of 3.5%, which was down from 3.9% last year…So I don't know that I can give you an answer that's more satisfying than last quarter on projecting what Apple phone sales might be. You've seen the numbers, they're down slightly over last year's levels on an introduction. We're still waiting, obviously, for the software release and whether or not that software release drives interest in the consumer base to accelerate that remains to be seen…I think some of these things are going to be a little bit more graceful ramp-up in consumer interest as opposed to a big bang.

Verizon Communications Inc. 3Q24 Earnings Call:

…we're not seeing a big upgrade cycle right now. The upgrades were down 10%. Right now, customers are choosing to hang on to their phones a lot longer, and that's by choice. The average upgrade rate [is] probably 40 months or so…I mean many of the AI application, of course, are very helpful. But when it comes to consumer devices, we also need to think about the processing power for this application, if you want to do something really innovative…I usually say, historically, when we've seen sort of cycles in our industry that's been 4G to 5G or hardware redesign…Now we're talking about is a software cycle's going to do it. It's too early for us to say at least, we're so far, and I'm looking at my colleagues, we haven't seen that. We haven't seen that is creating the cycle.

T-Mobile US, Inc. 3Q24 Earnings Call:

I mean in terms of the current upgrade rates, like you said they were low…So what we see is customers having devices that generally work better and devices that are becoming more expensive and lasting longer. And we think all those factors have come into play affecting and you are seeing the outcomes of the upgrade rates…as I think about Q4, we'll probably see the same seasonality from an equipment revenue perspective that we saw last year. So the same kind of absolute dollar increase that we saw Q3 to Q4 is, what I expect to happen from Q3 to this Q4.

Recent operating results from Apple’s largest global supply chain partners, including Foxconn and TSMC, have also shown early indicators of slowing iPhone uptake, let alone an outsized upgrade cycle. Foxconn (Apple’s key assembly partner) is now expecting AI server sales to eclipse iPhone assembly sales this year, despite the latter having historically been the manufacturer’s largest business segment. Meanwhile, TSMC (Apple’s chip manufacturer) has also been consistently reporting a diminishing share of smartphone revenues in the last two years.

These combined observations from Apple’s key industry partners give weight to the recent slew of weak global and China iPhone sales data for the December quarter. And the set-up is accordingly harbingering incremental downside risks of underperformance for Apple, which would bode unfavourably with the slow March quarter guide due to inevitable seasonal weakness.

However, current Wall Street estimates are still predicting outsized full year results that encapsulate expectations for an AI upgrade super cycle. Industry had previously estimated that a third of the 300 million iPhone 12 installed base that’s ready for upgrades is currently concentrated in China. But that optimism’s evidently stale now, given the sizable market share loss that the iPhone’s suffered in the region over the past year.

While some may blame Apple’s China weakness to the region’s mixed macro backdrop, demand dynamics for rivalling offerings are showing that the iOS exodus in the region is actually becoming increasingly structural. This is proven in the relentless demand for Huawei’s Mate XT trifold. Despite it being the world’s most expensive mass market smartphone, the Mate XT had garnered a waitlist of more than six million buyers leading up to its launch. Many are also still struggling to get their hands on the device today even if they’re willing to pay the hefty price. A similar trend’s also been observed in Huawei’s newest Mate 70 smartphone, which is the closer rival to the iPhone.

Assuming Apple’s more profitable services revenue stream continues to lead its growth in the double-digits percentage range going forward. And Mac, iPad, and wearables and accessories maintain a similar forward growth trajectory as historical levels. The iPhone would need a multi-year double-digit CAGR to support Apple’s current price of $223 per share:

And there are only a few scenarios where this upside scenario would materialize, yet none are feasible:

1. Apple benefits from AI-driven smartphone TAM expansion, while holding the iPhone ASP unchanged: This isn’t going to happen, since Apple Intelligence’s already proven to be a bust in spurring upgrades (both at home in the US and in China), let alone encouraging switches from non-Apple users to the iPhone.

2. Apple instigates a double-digit percentage increase to the iPhone ASP without compromising demand: Also unlikely, especially given the mixed global macroeconomic outlook and cheaper priced rivals in China that’ve already been eating Apple’s lunch.

3. Apple eliminates SE, non-Pro and low GB Pro options to artificially hike ASP: This would lead to the same conclusion as scenario (2). By eliminating the currently less expensive and lower-performance iPhone offerings, Apple risks further limiting its installed base by excluding price sensitive mass market consumers. This would also impact overall services uptake in the longer-term, leading to an adverse domino effect for Apple.

Looking ahead, Apple lacks new additive growth catalysts to support the stock’s valuation premium that’s been ballooning since June. Considering entrenched market share loss in China as the base case, and absent new growth catalysts, Apple’s intrinsic value veers closer to $200 per share.

The analysis estimates Apple’s net income to expand at an 8% five-year CAGR, driven by total net sales at a 5% CAGR over the same period. Ensuing free cash flows, inclusive of a terminal value estimated at a perpetual growth rate of 3.5%, are discounted at an 8.7% WACC. The perpetual growth rate considered is consistent with the estimated pace of expansion across Apple’s core operating regions, and reflective of Apple’s anticipated pace of steady-state growth in the long-run. Meanwhile, the discount rate applied in the analysis is consistent with Apple’s current capital structure and risk profile as well.

In the downside scenario, Apple could see a further entrenched “iPhone winter” in China, considering signs of large-scale user exodus to local rivals. This would also risk a spillover impact to services revenue growth, as the segment is largely dependent on the global iPhone installed base. The impact of margins would be equally detrimental as well. Slower uptake of Apple devices will trigger further discounts to encourage sales, leading to a reduced ASP. It’d also impact the longer-term growth trajectory for Apple’s currently most profitable and fastest growing services revenue stream, thus leading to further margin deterioration. This could easily take the Apple stock back to the $180-range that was observed prior to the introduction of Apple Intelligence during WWDC 2024, putting a full stop to the AI upgrade super cycle narrative.

Risks to Consider for Apple Bears

Shorting Apple is no easy feat. Apple’s still a great company with a loyal fanbase after all, thanks to its simple-to-use and seamlessly integrated ecosystem. The company also boasts one of the strongest balance sheets with a robust buyback program in place. This has, time and again, injected durability to the stock against wide swings driven by broader market volatility.

More importantly, Apple’s robust checkbook also makes its stock an attractive haven trade whenever market sentiment turns risk-off. This likely explains Apple’s rally observed during the final weeks of 2024. The adverse headlines pointing to risks of slowing growth at Apple at the time were likely overshadowed by market’s anticipation of a reduced scale of rate cuts heading into the new year instead. Yet shortly after stepping into the new year, Apple had relinquished the gains upon the release of a softer-than-expected December CPI report.

This suggests that Apple’s valuation is currently heavily influenced by macroeconomic dynamics. But this doesn’t discount the fact that Apple’s underlying fundamentals are weakening. And when Apple’s operating results regain precedence in dictating the stock’s valuation (which I expect to return post-earnings later this month), it’d give the short catalysts (i.e. F1Q25 earnings underperformance and nominal Apple Intelligence uptake in China) incremental weight.

Final Takeaways

Market confidence in Apple’s AI upgrade super cycle is rapidly fizzling, and urgency for management to come up with new growth catalysts is only increasing. What this means for now is that the risks remain skewed to the downside for Apple. This is evident in the Vision Pro’s failure in generating incremental growth momentum for the company, which was only made worse when Apple Intelligence debuted with a whimper in reaccelerating iPhone sales.

Despite Apple being a mainstay winner amidst macro-driven market volatility, the underlying business’ deteriorating growth outlook and earnings quality shouldn’t be dismissed – especially as fundamentals always take precedence in driving a stock’s ultimate valuation outlook.

This Apple analysis marks my first fundamental trade idea deep-dive on Substack — and it certainly won’t be the last, so please subscribe for follow-ups on Apple and other future deep-dives. I’d value any feedback as I continue to experiment with sharing on this platform. Thank you for reading.

Disclaimer: This analysis is for informational purposes only and represents the opinions of Livy Research. It is not investment advice nor a recommendation to buy or sell the securities discussed.

Remember, it’s not SMCI’s first clash with accounting-related filing delays and Nasdaq delisting risks. It’s happened before in 2017 – the whole thing dragged on for more than a year before Nasdaq bid SMCI goodbye and delisted the stock in 2019.

But SMCI didn’t have as much at stake at the time – it wasn’t a high-profile AI server provider, nor was it a S&P 500 company with substantial trading volume like it is today. Delisting risks are also high on the list of troubles facing SMCI today because of adverse liquidity implications tied to such an event if it were to happen. There’s an early cash redemption clause on the company’s $1.7 billion convertible notes that’d be triggered if the stock gets booted from Nasdaq.

Most importantly, SMCI’s latest capital raising efforts may be a telling tale that it won’t file in time and is heading for Nasdaq delisting. And CEO Charles Liang’s probably shot himself in the foot when he said otherwise just about a week ago.

The SMCI stock’s staged a nice recovery since the hiring of BDO and receipt of Nasdaq’s filing extension to February. But that’s likely the last hurrah before SMCI’s headed into another delisting spiral in the coming months.

SMCI’s Faces Imminent Delisting with Recent Fundraising Efforts

The trends observed to date show SMCI doesn’t take pre-emptive measures lightly. It’s not a big fan of taking mitigative actions as a “just in case” until it’s affirmative that the troubles waiting on the other side will happen.

For instance, SMCI only went on to prepay the term loans and other outstanding company debt (one after another) when it became evident the company wasn’t going to meet the renegotiated filing deadlines with the banks. And neither did SMCI pursue negotiation with the convertible note holders on a waiver clause for the early redemption in the event of a delisting – likely because it was confident in receiving the 180-day Nasdaq filing extension.

This implies SMCI’s latest capital raising efforts are in direct preparation for another round of filing delinquency in February. And management knows Nasdaq won’t be dancing around on the matter with them like it did in 2017.

The latest capital raising efforts represent a pre-emptive move by SMCI to mitigate post-delisting liquidity risks. Its pursuit of a Private Investment in Public Equity – or “PIPE” – is also a telling tale, since this arrangement can bypass disruptions from a potential Nasdaq delisting. A PIPE placement would give SMCI the wiggle room it needs with regards to filing deadlines and Nasdaq listing status, as long as the investors agree; this is a direct opposite from requirements under traditional fundraising through public markets. A PIPE also faces less regulatory hurdles with the SEC and grants the issuer faster access to capital. This is a critical consideration for SMCI, given the new Nasdaq filing deadline comes in less than two months with extended holidays in between.

The downside of an imminent PIPE placement is that existing SMCI investors will face the same dilution risks as if the company’s raising money through a public equity offering. Worst off, PIPE shares are also typically sold at a discount to market value to compensate for incremental risks associated with the investment – which is exactly SMCI’s case. This arrangement effectively heralds significant volatility to the stock, which SMCI’s already not in a shortage of given overhanging delisting risks.

How Close is SMCI from Being Delisted?

One might argue that CEO Charles Liang has only just given his word that SMCI will file by February 25 and not be delisted from the Nasdaq. But SMCI’s got a track record of making these empty promises – just like it did when it most recently renegotiated filing deadlines with the banks. What matters really is the action SMCI’s taken amidst the situation. This accordingly suggests that SMCI’s pre-emptive capital raising efforts and specific focus on the PIPE route is a direct response to an imminent filing delinquency and subsequent Nasdaq delisting.

This is further supported by the fact that Nasdaq’s unlikely to be as lenient as it did when SMCI repeatedly failed to file the FY17 10K and subsequent quarterly statements in time. So far, things have largely unfolded the same way it did in 2017 – SMCI’s delinquent on the FY24 10K due to accounting-related issues, and Nasdaq’s granted the first 180-day filing deadline extension.

But SMCI’s also dealing with a brand-new auditor this time to assess its books – this is an incremental challenge. A new audit engagement would require the auditor to perform additional work on prior filings before starting on the existing audit. This includes review procedures on FY23 statements to ensure the FY24 opening balance can be relied on when performing the audit. On top of that, there are additional audit procedures that’ll need to be performed given issues on the effectiveness of internal controls as cited by SMCI in its 10K filing delay notice. And some of these could potentially be tasks that BDO and SMCI won’t be able address by the February filing deadline even if there’s additional staffing working round the clock. To have the FY24 10K audited without issue by February 25 is an almost impossible task to fathom for an experienced auditor on a rolling engagement, and let alone a newly engaged one like in the case of BDO and SMCI.

To make matters worse, SMCI’s also officially been kicked out of the Nasdaq 100. It was earlier speculated that Nasdaq’s granted the initial 180-day filing extension to SMCI to buy time for index funds to regroup themselves. With that now out of the way, and Nasdaq’s knowledge of the immediate audit challenges facing SMCI, there’s one less reason to keep the stock listed on the exchange. It’s really no coincidence that word of SMCI’s capital raising efforts were reported around the same time it was booted off the Nasdaq 100.

What Happens When SMCI’s Delisted?

The SMCI delisting will take some time to materialize – even when it becomes delinquent with its filing on February 25. The last time, it took about seven months between the time when SMCI exhausted Nasdaq’s maximum 360-day filing extension (August 2018) and when the stock became officially delisted (March 2019).

But given SMCI’s current situation and history with filing delinquencies, it’s unlikely Nasdaq will give it the full 360-day extension (nor additional days of listing stay in between).

Once SMCI confirms it won’t file by the initial 180-day filing extension on February 25, Nasdaq will give it one week to request a hearing from the Hearings Panel. When this request is put through, SMCI is automatically granted a 15-day listing stay that begins after the hearing request deadline. SMCI can request for the 15-day listing stay to be extended until the Hearings Panel’s made a decision on whether to grant another filing extension – which is what it did successfully in 2018. If the Hearings Panel rejects SMCI’s plea for another extension, the company can still appeal to the Listing Council.

And if that fails too, then Nasdaq will notify SMCI of the stock delisting process, which takes about 10 calendar days before the name’s officially removed from the exchange. Prior to delisting, the stock can also be subject to suspended trading in the process of ongoing hearings and appeals. SMCI dropped 15% overnight when this happened in August 2018, highlighting the heightened volatility it’s subjected to amidst the process.

Once Nasdaq confirms SMCI’s delisting from the exchange, the early cash redemption clause on the $1.7 billion convertibles will be triggered. Based on SMCI’s current cash position, it likely won’t have the money to repay the convertible note holders even if a delisting doesn’t take place until late 2025.

SMCI’s latest “business update” reported $2.1 billion cash on hand as of September 30, 2024. Since then, it’s repaid $500+ in term loans, and an undisclosed amount related to outstanding lines of credit. Its ability to generate cash flows from operations also remain uncertain, given recent speculation on a string of “client exodus” in response to SMCI’s accounting woes.

This leaves SMCI’s ultimate liquidity stance up to its current PIPE capital raising efforts, which we believe will be sized according to the convertible redemptions and other working capital needs. If the PIPE funds can cover SMCI’s impending cash needs, then the company would be safe for now from an operating perspective.

However, the picture’s comparatively bleak from a valuation perspective. Once SMCI’s delisted, it’ll trade over-the-counter (“OTC”). Institutional stake in the stock would’ve significantly decreased by then. This is in line with current portfolio rebalancing taking place in response to SMCI’s removal from Nasdaq 100, and similar undertakings expected when the stock becomes delisted and removed from the S&P 500 as well. It's likely executive level management has also collateralized their respective stakes in the company for personal reasons. And these internally collateralized stakes could further fuel volatility when combined with institutional selling. Trading OTC is also significantly less liquid than public markets due to reduced institutional exposure and the cohort’s inherently elevated investment risks.

The last time SMCI was delisted, its share price had remained largely stable in OTC trading. But this was likely because its market cap was much smaller with limited institutional exposure at the time. However, there’s more at stake this time around given the prominence that SMCI’s technologies have risen to in recent years and the ensuing popularity it’s gained with both institutional and retail investors alike. A Nasdaq delisting and subsequent OTC trading will wreak havoc for the stock, even if subsequent liquidity risks are mitigated.

Key Watch Items Ahead

The first order of business would be to keep an eye out on the size and pricing of SMCI’s upcoming PIPE investment consortium. The closer the amount raised is to the $1.7 billion outstanding convertibles, the more likely SMCI’s preparing for filing delays and a subsequent Nasdaq delisting.

Developments leading up to SMCI’s extended February filing deadline are also crucial. They will underpin increased volatility to the stock, similar to the weeks leading up to Nasdaq’s initial filing extension decision in early December. In the event SMCI fails to file by February 25, markets should expect substantial institutional selling at the minimum, given the elevated risks of delisting that SMCI faces.

It's also possible that SMCI won’t be officially removed from the Nasdaq until later 2025 even if it doesn’t file by the February deadline. This is consistent with the administrative process of appeals and delisting-related paperwork as discussed earlier. Under such circumstances, SMCI’s accounting debacle would overlap with the new year’s PCAOB inspection results. And BDO’s high up on the list of troubled public accounting firms. This means any new inspection findings on BDO audits will directly impact the SMCI stock’s performance, as investors respond to implications on the company’s outstanding filings and Nasdaq listing status.

The PCAOB typically publishes their inspection reports for audit firms with findings in May. And BDO gets unusually high focus this year – not only because it’s been in the lead for inspection findings and material deficiencies by wide margins in recent years, but also because it’s been singled out by the Senate as the problem child.

In a letter dated October 9, 2024 to the PCAOB, the Senate’s specifically called out the integrity of BDO audits. The letter points to BDO’s 2023 audit deficiency rate, which reached an all-time high of 86%, ballooning from 66% in 2022 and 53% in 2021. BDO’s audit deficiency rates also lead the next-worst player, Grant Thornton, by 32 percentage points.

Given BDO’s decreasing audit quality, combined with historical penalties from the PCAOB that’ve been merely viewed as a “hand slap”, the Senate is now calling for greater scrutiny on the firm’s work.

This means the upcoming round of PCAOB inspection reports, particularly on BDO audits, will matter for SMCI. The audit firm’s tie-up with SMCI’s high-profile accounting saga doesn’t help with the airtime either.

BDO’s inspection results will directly dictate whether the eventual SMCI filing will be subject to another round of regulator inspections. If SMCI’s outstanding 10K and subsequent 10Qs are filed by February, then it’ll likely be subject to PCAOB inspection for the 2024 audit year. And if SMCI’s outstanding filings aren’t completed by early 2025, then the subsequent 2024 inspection report for BDO will cast a darker shadow over the validity of SMCI’s audit amid its complex accounting backdrop. And neither situations are beneficial to prospects of the SMCI stock.

Final Takeaways

SMCI’s been paring some of its losses since engaging BDO as its external auditor and receiving the filing extension from Nasdaq. But dismissing its delisting risks altogether would be a mistake.

My bearish outlook on SMCI is not so much about whether the company’s committed fraudulent reporting, but rather the concerns over its repeated accounting shortfalls and the subsequent implications of an imminent Nasdaq delisting. Given SMCI’s latest update on capital raising activities through a PIPE, I’m almost certain the company won’t be filing in time and is preparing to be booted off the Nasdaq.

And the troubles won’t end there, even if SMCI manages to dodge a liquidity spiral related to the early cash redemption on its outstanding convertibles. SMCI will likely face protracted uncertainties to its fundamental outlook, as suppliers and customers adjust for inevitable exposure to the sever maker’s unresolved accounting issues. This could very well lead to a further selloff in OTC trading as the company’s valuation corrects for related fundamental shortfalls. Reduced liquidity in OTC markets could represent an additional multiple compression risk for SMCI, which shouldn’t be overlooked.

This SMCI analysis marks my first catalyst-driven trade idea analysis on Substack. I’d value any feedback as I continue to experiment with the kind of information I’m hoping to share on this platform. Thank you for reading.

Disclaimer: This analysis is for informational purposes only and represents the opinions of Livy Research. It is not investment advice nor a recommendation to buy or sell the securities discussed.