Nomura research reveals the fading of "AI growth panic"! Profit realization and capital inflow resonate, with the AI computing power theme leading technology stocks' rebound

According to The Smart Investor APP, Nomura cross-asset strategist Charlie McElligott's latest report indicates that the AI trading theme is emerging from a previous round of panic over AI computing power and technology growth, which was amplified by concerns over surging AI financing costs, overcrowded positions, and accelerated deleveraging by hedge funds.

This top Wall Street strategist from Nomura emphasizes that large-scale AI financing has found buyers, semiconductor giants and the two major cloud computing titans have generally delivered strong results, and Anthropic's IPO filing disclosures have collectively eased the persistent "AI growth panic" of the past two weeks. The unwinding of crowded and over-leveraged positions has also set the stage for a market recovery.

Large financing deals have found backing, semiconductor giants closely linked to the AI compute infrastructure process—as well as the cloud leaders Microsoft and Alphabet—have reported impressive earnings, and Anthropic’s updated IPO filing prompted the market to reassess the previously pessimistic narrative that "rising funding costs will inevitably hinder AI expansion." At the time of McElligott’s report, Nasdaq-related volatility dropped significantly for three consecutive trading days; the long-short pair trade of "long AI infrastructure beneficiaries, short hyperscale cloud vendors" is still down 11% for the month but has rebounded 8% in the past week—Nomura points out this highlights investors beginning to recognize the profitability of upstream suppliers again, correcting the valuations previously suppressed by forced deleveraging and a rush to reduce crowded trades.

Meta’s equity financing, equipment financing for entities related to Broadcom and SpaceX, addressing "who will foot the bill for compute expansion," and subsequent reports of 48% growth in Microsoft Azure and 103% growth in Alphabet Google Cloud—together with strong profits at SK Hynix—have all collectively validated the commercial realization of AI demand. As volatility in tech stocks drops and credit spreads for some AI firms narrow, the market has begun to reevaluate growth prospects amid funding pressures. The AI computing theme is thus supported by both capital and earnings growth, driving a further strengthening of the semiconductor-led tech stock rebound narrative.

AI Trading "Cleanup" Relieves Pressure! Nomura Signals End of "Two-Week AI Growth and Credit Panic"

Nomura cross-asset strategist Charlie McElligott noted that after a brutal two-week sell-off, the global equity markets have finally breathed a sigh of relief: a wave of financing in the global technology sector, the far stronger-than-expected earnings from semiconductor leaders focused on AI computing infrastructure, and Anthropic’s comprehensive S-1 IPO filing have collectively shattered the bearish narrative that the investment thesis for AI compute infrastructure has completely collapsed.

This relief is particularly evident in the derivatives trading division. Over the past three sessions, volatility in Nasdaq-related assets dropped sharply from last week's "extreme tightness," as described by McElligott. Meanwhile, the previously battered trade—long AI infrastructure beneficiaries, short hyperscale cloud vendors—is still down 11% month-to-date, but rebounded strongly by about 8% in the last week alone. This shift comes as earnings from Alphabet (GOOG.US)(GOOGL.US), Microsoft (MSFT.US), and Amazon (AMZN.US), as well as results from semiconductor giants like Samsung, Intel (INTC.US), and AMD (AMD.US), become the market’s focus. Recent earnings from Microsoft, Alphabet, and last week’s report from Facebook’s parent Meta have all powerfully demonstrated the clearer monetization path and strong revenue associated with AI-driven applications.

In his client note, McElligott wrote that the IPO filing from global AI powerhouse Anthropic (ANT.US) on Monday "has played an extremely important role" in restoring confidence in the expansion of AI compute infrastructure demand and the economic viability of AI.

The capital markets are also reinforcing this investment thesis. McElligott pointed out that Meta’s planned equity issuance, which will ultimately reach $98 billion, has seen its initial offering absorbed by the market; Broadcom and Anthropic have struck a $63 billion chip deal; SpaceX and Apollo Asset Management struck a $40 billion equipment financing lease agreement. Overnight, SK Hynix, the world’s leading storage chip company, released stellar quarterly results—revenue up 328% year-on-year, operating profit margin hitting 80%, and management announced capacity “sold out until 2029”—pushing its stock price up double digits.

The strategist sees the recent momentum reversal and portfolio liquidations at Pod-driven hedge funds as a natural consequence of overcrowded positions. He notes that multi-strategy funds remain the main destination for allocators, but high leverage and persistent crowding risks lead to cyclical cleanups.

"All of this, too, shall pass," McElligott wrote, "and then changes in performance will play a positive role, clearing excessive positions and leverage out of the system."

From "Who Pays the Bill" to Capital Re-engagement: Material Relief in AI Financing Pressure

One of the most critical changes behind this rebound is that capital markets have begun to provide executable financing arrangements for AI expansion. For example, McElligott cited Meta’s first batch of a planned $98 billion equity financing being taken up, Broadcom’s $63 billion chip financing for Anthropic, and the $40 billion equipment leasing arrangement between SpaceX and Apollo. These three deals, through equity, asset financing, and leasing, respectively, help reduce the pressure on companies to raise all construction capital at once.

Among them, the Broadcom deal uses a special purpose vehicle (SPV) to purchase AI accelerators and network equipment and then leases them to Anthropic, with delivery expected to start in 2027. The most direct positive impact on the semiconductor and AI computing power supply chain is the improved ability to convert client procurement budgets into equipment delivery and supplier revenue.

The credit market is also responding. Markit data shows that CoreWeave’s 5-year credit default swap (CDS) spread narrowed to 1,041.4 basis points intraday on October 27, a more than weeklong low. According to the latest CNBC coverage, Meta’s CDS spread also contracted after it announced its equity financing. The basic underlying mechanism here is clear—increasing equity capital and reducing market expectations for imminent large bond issuances simultaneously improve the capital structure and supply-demand in credit markets.

Therefore, the more worthy trend now is that while financing costs remain high, the accessibility of capital is improving, and market concerns about AI projects being forced to downsize are cooling off. This also explains why equities can start pricing in order and profit recovery first, while bonds of companies with varying creditworthiness still show clear divergence.

With Wall Street giant Goldman Sachs landing a $210 billion financing mandate, the U.S. semiconductor sector extended its rally from last Friday and the market started to price in the "capital relay" logic, Wall Street institutional investors’ worries about an AI compute infrastructure funding crisis have notably eased.

Last week’s intense sell-off in U.S. semiconductor and broader technology stocks was driven by market participants' collective questioning of "who will bear the massive, trillion-dollar costs of AI compute infrastructure construction." The latest announcements partially answer Wall Street’s doubts: Broadcom and Anthropic’s eventual $63 billion AI chip deal, Meta’s equity financing, SpaceX’s leasing of large-scale AI compute resources, and Anthropic’s IPO funding all underscore that capital expenditures for AI are likely to continue receiving strong market support. Consequently, the market is reassessing the likelihood that funding pressures could slow down AI computing resource procurement or delay large orders for AI chips and other core infrastructure hardware—this "capital relay" mechanism is at the heart of the ongoing global semiconductor sector’s rebound.

Market action provides direct evidence: The Philadelphia Semiconductor Index, after plummeting 6.72% on October 22, rebounded for four straight sessions and closed at 11,684 points on the 28th, up about 8.87% since the 22nd; the Nasdaq 100 rose for three consecutive days from the 26th to the 28th, rebounding by about 2.95%. Financing activities and storage sector results first stabilized the market, then the post-market earnings reports from Microsoft and Alphabet on the 28th further substantiated this recovery with cloud business growth evidence. Mechanically, crowded position sell-offs easily trigger risk limits, stop losses, and synchronized reduction in holdings; when fresh good news improves profit expectations and volatility drops, forced selling pressure can ease, and fundamentals once again become a key pricing factor.

Cloud Revenue and Storage Profits Validate Each Other: "AI Conviction" Now Strongly Supported by Real AI Application Results

Strong growth in the cloud businesses at Microsoft and Alphabet, and the string of positive growth updates from AI compute industry leaders such as SK Hynix, Seagate, and TSMC, have provided substantial commercial evidence to bolster Nomura’s call for a sentiment recovery.

Microsoft Azure cloud revenue jumped 48% year-on-year; total Microsoft Cloud revenue hit $63.9 billion, up 30%. Management expects Azure to grow 49%–50% next quarter, with free cash flow for the quarter of about $18.6 billion—much better than market expectations. Alphabet’s third-quarter cloud division—Google Cloud—recorded $30.709 billion in revenue, up 103%; growth was driven by enterprise AI infrastructure, AI solutions, and core cloud services, with TPU system sales being an important incremental driver. The joint takeaway from the two earnings reports: AI investments are now generating measurable business revenue, and Microsoft’s cash performance has also strengthened confidence in the large cloud providers’ capacity for sustained investment.

Upstream, there’s a cross-validation of profits and orders. HBM and server DRAM leader SK Hynix reported Q3 revenue of KRW 104.62 trillion, up 328% year-on-year; operating profit hit KRW 83.33 trillion, up 632% year-on-year, with an operating margin of 80%. Its long-term agreements with 14 major clients now cover more than half of 2027–2028 capacity; management’s view that capacity is "substantially sold out" through 2029 also factors in existing agreements and current negotiations.

From an industrial mechanism perspective, as intelligent agents execute more tasks, the demand for model inference, tool invocation, context processing, and data retention rises—driving increased consumption of computing, memory, storage, and connectivity resources. When such demand translates into cloud revenue, long-term procurement commitments, and supplier profits, the semiconductor bull cycle is supported by much more than just rising capital expenditure. What truly reinforces the resilience of the AI trade is the mutually reinforcing cycle between rapid AI commercialization around cloud business, continuous capital engagement, and profitable scarce supply.

Furthermore, bullish Wall Street perspectives are converging from different angles. Goldman Sachs’ "GOAL Kickstart" maintains an overweight on equities for 3- and 12-month horizons even as risk appetite metrics drop to around -0.5, citing earnings resilience, lower valuations, and reduced crowding. Recent tech stock portfolio reductions have created a more flexible environment for trading in new earnings positives. Morgan Stanley’s Wilson maintains an 8,000-point year-end target for the S&P 500, listing AI financing, the midterm elections, and Federal Reserve liquidity as key catalysts for an adjustment’s end; the absorption of new financing directly addresses the issue of capital supply. Deutsche Bank provides a seasonal perspective: in 21 of the last 23 midterm election years, the market rose in the three months following early October. Each of these views addresses valuation and positioning, event digestion, and historical seasonality, while the latest robust results in cloud and storage add tangible earnings support to the year-end bullish narrative.

The rebound narrative for the storage sector is even more focused on the gap between supply scarcity and pessimistic valuations. Goldman Sachs has lowered its KOSPI target from 12,000 to 9,000, but the current level still implies roughly 62% upside, and Goldman argues that the implied scenario of a 65% earnings decline for Korean semiconductor stocks next year is overly pessimistic. Meanwhile, at a TrendForce expert meeting organized by Goldman, the consensus was that DRAM price hikes could last through the first half of 2028, while HBM’s blended average price could rise 122% year-on-year in 2027; clean room, equipment delivery, and product upgrade constraints make it difficult for new investment to immediately turn into effective supply. Reports that Alibaba, Tencent, and ByteDance are seeking long-term DRAM contracts also indicate that demand for secured supply is spreading to more global-scale cloud providers.

Capital has already started to optimistically bet on an AI trading frenzy comeback: Between October 1 and 23, Korean retail investors bought leveraged products related to Samsung and SK Hynix worth KRW 315.4 billion; by October 22, Korean investors had bought $3.548 billion of SOXL (the 3x leveraged ETF tracking the Philadelphia Semiconductor Index, US ticker “SOXL”), amounting to 73.9% of Korean investors’ net US stock purchases in the period. The core of the storage sector’s rerating is the visibility of orders and profit performance exceeding the contraction scenario previously priced into stocks, with capital inflows now amplifying revisions to growth and expansion expectations.