Neutral
ASML
IRENIREN LIMITEDIREN LIMITED
TSMTSMCTSMC Weekly|Anthropic ARR, Neocloud Credit Risks, TSM & ASML, Muse Wonder, Physical AI, Enterprise AI Spending, BE
The AI trade has become much more sensitive to near-term numbers. As we argued in our latest column, investors are watching ARR, personal-agent DAU and retention more closely, while higher yields have become a headwind. A compelling long-term story now has to contend with monthly growth data, and the recent plateau in personal-agent usage has made that shift more uncomfortable. Credit markets added another source of pressure this week. CoreWeave’s bond yields climbed to around 15%, and its five-year CDS reached roughly 1,195bp on Thursday, as investors questioned the funding behind the AI buildout; Nebius and IREN also sold off. For leveraged neoclouds, slower revenue growth becomes more damaging when fresh capital is expensive or harder to obtain. That can delay data-center expansion and chip purchases, spreading the concern across the AI supply chain. We see the financing squeeze as an important amplifier of the selloff: the market is now testing both the revenue trajectory and the balance sheets needed to fund it. We flagged a related demand-side risk in our September 2 Weekly Call with Premium Subscribers . Newly funded AI labs, or Neolabs, had become important marginal buyers of short-term GPU capacity, paying high rental rates despite uncertain business models. We noted that their funding could support roughly six to twelve months of spending, after which continued purchases would depend on raising more capital. We are watching this closely: if risk appetite weakens and the next funding round becomes harder, these labs could cut compute purchases, putting pressure on neoclouds’ short-term rental rates. Financing conditions can therefore affect both neoclouds’ cost of capital and their customers’ ability to pay. Anthropic’s $16.8b of 3Q26 revenue puts September exit ARR at about $70b in our model, below the $80b expectation we had been tracking. Giving up most distillation revenue explains part of the shortfall, but the harder question is the path to $100b by year-end, a market expectation that we do not believe the company has explicitly guided to. Our model puts September trailing-28-day ARR at about $69b; reaching $100b would require roughly $9b of monthly net new ARR, against about $5b excluding distillation over the past quarter. Opus 5.5 could help growth accelerate, but gains at Anthropic may then renew concerns about OpenAI’s growth as the competitive debate intensifies. We think year-end revenue delivery will keep driving the AI trade, even after the pullback. Expectations have become harder to judge, and actually reaching $100b could still surprise the market. We are looking for the next use cases that can broaden revenue growth, from drug discovery and diagnosis to supply-chain automation and cyber defense. Getting closer to AGI may make the investment experience harder along the way: more compute goes into training before the return is easy to demonstrate. Finding the next large application before that turning point remains central to our research. This Week’s Reports Subscribe now Read more
View original →
ASMLNeutralReview|ASML 26Q3: Upgrades Lift Gross Margin; Capacity Expansion Awaits Further Clarity
Customer prepayments surge as demand signals extend into 2030 Customer prepayments surge as demand signals extend into 2030
Summary
View original →
ASMLNeutral
MRVLMarvellBullishMarvell
OCSOculis HoldingBullishOculis Holding
CRWVCoreWeave, Inc.NeutralCoreWeave, Inc.
GOOGLAlphabetNeutralAlphabet
METANeutral
NVDANVIDIANeutralNVIDIA
OPOptimismNeutralOptimism
TSMTSMCNeutralTSMCWeekly|AI4S, Gemini 4, TSM & ASML, MRVL and OCS, Muse, CRWV
Google is back in the frontier race, and we think a new training cycle is beginning. As we argue in this week’s column, Gemini 4 is evidence that scaling still works; the next test is how much post-training improves coding and long tasks in Gemini 4.1. With OpenAI, Anthropic, Google, Meta and xAI all investing to defend or improve their positions, we remain bullish on the networking and optical interconnects needed to keep larger clusters productive. AI for science gives us another reason to be optimistic about where that compute can go. OpenAI’s latest mathematics release still needs independent scrutiny, but progress in problems with verifiable answers could help AI take on more research and engineering work. Drug discovery, materials and energy research offer much broader potential benefits. Physical experiments and clinical trials will still take time; our optimism is about expanding the ideas scientists can test and the connections they can make across disciplines. The supply chain is preparing for that investment. We see TSMC’s growth increasingly governed by how quickly capacity comes online, and have raised our ASML EUV shipment forecasts again. Marvell’s Investor Day targets are moving closer to our forecasts, backed by longer-term supplier planning. OCS adds another source of optical demand as content per accelerator rises, although deployment beyond Google remains an important assumption in our longer-term model. On the application side, we are watching retention and cost together. Muse’s improving retention gives Meta a reason to expand acquisition once the economics support it, while computer use opens a much larger pool of work to automation. Our expert interviews also underline the obstacles: high token consumption can limit enterprise adoption, and physical capacity still depends on manufacturing yields, equipment and power. Better models and more infrastructure need to translate into reliable, affordable work. This Week’s Reports Going the Way of Go: stronger models make the next training cycle worth watching. Our new column connects AI’s progress in mathematics and programming with Google’s return to the frontier, and explains why we remain optimistic about AI for science and optical interconnect demand. TSM & ASML: capacity expansion sets the pace. We expect TSMC’s advanced-node expansion to support further growth and now forecast ASML total EUV shipments of 100/125 units in 2027/28, as both logic and memory customers need more equipment. MRVL & OCS: more optical content as AI clusters grow. Marvell’s new targets move closer to our forecasts, while our OCS work points to more than $20b of annual demand by 2030, conditional on broader adoption that includes Nvidia from late 2028. META: Muse’s retention is improving before the next growth push. We expect adoption to advance in steps as Meta improves retention and balances user acquisition against the substantial compute cost of serving a personal agent. Computer use: a much larger market for AI providers. Our long-term framework points to roughly $1tn of additional annual revenue and automatable work equivalent to about 180mn full-time roles, an estimate of task hours rather than a forecast of layoffs. Premium Report Snapshot Subscribe now Read more
View original →
ASMLNeutralPreview|ASML 26Q3: Raising EUV Shipment Expectations Again
Capacity Expansion Likely to Accelerate Capacity Expansion Likely to Accelerate
We believe ASML will further ease its capacity bottlenecks, with 2027 low-NA EUV capacity likely to exceed 95 units, above the company’s current guidance of 85 units. Our industry checks also indicate the supply chain is planning production around higher shipment targets.
View original →