CoreWeave Surges 25%, Igniting New Cloud Sector; Next Test for AI Compute Trading Is NVIDIA Earnings

Author: Long Yue, Wallstreetcn

CoreWeave released its Q3 earnings report on Tuesday, with both revenue and losses beating expectations, while its full-year guidance also exceeded the upper end of market expectations. The news boosted the "NeoCloud" sector, which had been under pressure due to concerns over AI investment returns.

Driven by CoreWeave's single-day stock surge of approximately 25%, the NeoCloud sector saw intraday gains exceed 12%, while the Nasdaq also recovered ground lost after the midterm elections.

Market optimism was further reinforced by two pieces of news: reports indicated Anthropic is discussing an IPO valuation exceeding $1.5 trillion, alleviating prior concerns about a significant downward revision in its valuation; additionally, OpenAI posted strong quarterly revenue, reaching $10 billion for the first time.

For Anthropic, which is set to enter the capital markets next week, this earnings report could not have come at a better time: CoreWeave is one of Anthropic's key compute providers, and the fates of the two companies are partially intertwined. Meanwhile, Nvidia's earnings report next week will serve as another litmus test, as investors will look for more reasons to judge whether the AI infrastructure boom can sustain.

However, the foundation of this rally remains shaky: on one hand, capital is shifting from hyperscale cloud providers to "shovel sellers" and compute leasing firms; on the other hand, CoreWeave's future funding gap, debt costs, and compute supply expansion remain the core variables determining whether this AI infrastructure trade can run its course.

CoreWeave Drives Collective Rally in NeoCloud Sector

CoreWeave reported Q3 revenue of $3.66 billion, surpassing analyst expectations; it recorded a loss of $1.27 per share, a narrower miss than the market had previously feared. More closely watched was its full-year guidance: the company expects 2026 full-year revenue to range between $13 billion and $13.3 billion, roughly 3.4% above the upper bound of consensus estimates.

During a call with analysts, CoreWeave CEO Michael Intrator revealed a noteworthy detail: more than half of this year's new customers come from outside the AI sector. This data partly addresses market concerns about the over-concentration of AI compute demand, which fears a sharp demand cliff should the AI investment frenzy cool.

Intrator's remarks suggest that demand for high-performance compute is penetrating broader industries, no longer relying solely on procurement by AI labs.

Prior to this, CoreWeave's stock had cumulatively fallen 39% year-to-date, making it one of the worst-performing stocks in the AI compute sector. Deutsche Bank analyst Brad Zelnick noted after the earnings release that he thanked the company for providing "reassuring details," a sentiment that largely captures the overall market mood.

Following the earnings announcement, CoreWeave's stock surged approximately 25% in a single day, becoming the primary driver of the NeoCloud sector's rebound.

Chart: CoreWeave (CRWV) 60-minute candlestick chart, rebounding to $59.78 (+22.68%) on November 11 after pulling back from its October high.

Market data showed that NeoCloud concept stocks rose over 12% as a group that day, while Goldman Sachs' broad AI stock basket gained approximately 5%, significantly outperforming non-AI equities. Previously weighed down by concerns over ROI on AI capex, the compute industry chain experienced a rapid valuation repair.

Chart: Comparison of the performance of Goldman Sachs US NeoCloud and other AI sub-sectors versus the S&P 500 index (excluding AI components) since November 2.

The Nasdaq also recovered ground lost after the midterm elections.

Chart: Intraday performance of major US equity index futures such as S&P 500 (@ES) and Nasdaq (@NQ), with Nasdaq futures leading the rally up 2.294%.

However, from a trading structure perspective, capital allocation reflects that investors have not fully dispelled their doubts. "Payees" of AI capex, such as compute leasing firms and semiconductor companies, outperformed, while hyperscale cloud providers like Microsoft, Google, and Amazon, which bear massive data center investments, lagged relatively behind.

In other words, the market currently prefers to bet on direct beneficiaries of the AI construction cycle rather than the "payers" who continue to foot the bill for massive capital expenditures.

Chart: Performance comparison between US hyperscale cloud providers ("payers") and the US AI semiconductor sector ("payees") since November 2.

CoreWeave and Anthropic: Intertwined Fates

CoreWeave's earnings report holds direct significance for Anthropic.

Anthropic is a customer of CoreWeave, utilizing compute power from CoreWeave and several other NeoCloud providers to support its Claude product line, while also holding contracts with hyperscale cloud vendors like Google.

Neither company has publicly disclosed contract terms, but Intrator described their collaboration in April as a "multi-billion-dollar agreement."

According to Bloomberg, Anthropic plans to complete its listing next week. As the first US frontier AI lab seeking a public offering, Anthropic lacks traditional moats like search engines, social media, or cloud businesses, meaning its valuation will hinge entirely on the market's assessment of AI prospects.

CoreWeave's robust performance theoretically helps Anthropic secure higher valuation recognition during its roadshow.

Recent market news further heightened interest in this IPO: reports indicate Anthropic's underwriters are evaluating the possibility of an IPO valuation of $1.5 trillion or higher.

If this valuation expectation holds, Anthropic will serve not only as a test case for the commercialization capabilities of frontier large model companies, but also as a direct barometer of the premium the market is willing to pay for AI's long-term growth story. Previously, there were rumors that Anthropic's valuation might be significantly discounted, falling below $1.5 trillion.

Meanwhile, the news that OpenAI's quarterly revenue reached $10 billion for the first time provides fresh support for AI application-layer demand.

For investors, this means the market is attempting to find a closed loop where "model revenue growth, compute procurement expansion, and returns on chip and data center investments" mutually validate each other; however, whether this loop is sufficient to justify high valuations remains to be verified by additional corporate earnings and order data.

For Anthropic, the positive signal released by CoreWeave's earnings lies in the fact that its upstream compute supplier continues to secure strong orders, raise revenue guidance, and maintain financing capabilities.

Yet this also highlights a shared risk: whether large model companies can sustain high growth in external compute demand, and whether compute suppliers can maintain stable delivery amid high capex and debt levels, will jointly determine the valuation foundation of this supply chain.

Notably, Anthropic relies on several other NeoCloud providers with unclear prospects for compute supply, including Nscale, which is reportedly considering delaying its own IPO plans.

The Next AI Test: Nvidia Earnings

Nvidia will release its earnings next week. If investors can temporarily set aside concerns about CEO Jensen Huang and related issues, Nvidia's results may provide the market with more reasons for optimism.

The importance of Nvidia's earnings report extends beyond verifying robust GPU demand. The market hopes to confirm three things: whether purchasing intensity from cloud providers and NeoCloud services remains steady, whether supply and delivery of new products like Blackwell can meet demand, and whether rapid growth in customer capex can translate into more sustainable revenue and profits.

During this rally, the semiconductor sector performed strongly, with the PHLX Semiconductor Index posting its largest single-day gain in months and recovering most of the losses from the previous week's pullback; Nvidia's stock also rebounded. However, traders observed that some active funds and retail investors still sold ahead of Nvidia's earnings.

This indicates that market expectations for the chip leader remain extremely high. If the earnings report fails to provide sufficiently strong guidance, volatility in AI trades could flare up once again.

Chart: Performance of the Goldman Sachs US AI Semiconductor Index (GS US AI Semis) from October to November and recent rebound trend.

Bloomberg columnist Dave Lee analyzed that back in March 2025 when CoreWeave went public, he warned that the company would become the "first true bellwether for the AI revolution," and events have proven him right: CoreWeave's stock price has become the market's primary "lightning rod" for volatility.

Behind the Rally: Funding Gaps and Supply Expansion Remain Hard Constraints

Nevertheless, the foundation of this rally remains fragile. CoreWeave's future funding gap, debt costs, and compute supply expansion remain the core variables determining whether this AI infrastructure trade can sustain.

CoreWeave's performance eased short-term concerns about whether AI construction would abruptly halt due to financing constraints, but it did not definitively answer long-term questions.

The bull thesis argues that CoreWeave can establish financing entities around customer contracts, using highly collateralized debt to cover construction costs; after repaying the debt within the contract period, the company retains the GPU assets and can re-lease the fully amortized equipment. If compute remains structurally scarce long-term, this model could generate continuous cash flow from existing inventory.

Sellers focus on three risks. First is pricing: there remains a gap between the per-megawatt revenue level mentioned by management and the actual monetization reflected in financial statements. Second is financing: market views suggest the company could face a funding gap of approximately $33 billion next year, still reliant on new debt or customer advance payments to fill it.

Finally is supply: before 2028, the market expects approximately 55 gigawatts of new compute capacity to come online. If supply release outpaces demand growth, rental rates and residual values for older generation GPUs could come under pressure.

Thus, what CoreWeave proved this quarter is that when compute remains tight, the company can secure orders and obtain financing; however, it has yet to prove that supply tightness will persist long enough to cover massive debt and support the residual values of older GPUs. Moving forward, the market will closely monitor the final pricing of its investment-grade secured loans, the scale of customer advance payments, and whether the inflection points for corporate capex and free cash flow materialize as expected.

Chart: CoreWeave Deferred Drawdown Term Loan (DDTL) and SPV financing operational workflow (Source: Goldman Sachs GIR).

Macro Environment Not Clear, Regulatory Uncertainty Remains a Hidden Risk

Following the US midterm elections, Democrats control Congress. Texas Governor-elect Gina Hinojosa advocated for a pause on data center construction during her campaign. Intrator wrote to Hinojosa last week, hoping to reach a consensus on "responsible completion" for approved projects and "enduring standards" for future ones.

During the earnings call, Intrator attempted to calm investor sentiment. He stated, "We do not expect to be affected by recent major regulatory changes. We view these changes as statements on how to build, rather than whether to build."

Beyond regulation, interest rates could remain another constraint on AI infrastructure valuations. During this stock market rally, market pricing for future rate hikes has risen, with 10-year US Treasury futures implied yields nearing 5% again. For NeoCloud companies reliant on debt-financed data center construction, rising financing costs will directly impact project returns and equity valuations.

More importantly, volatility across equities, credit, rates, and crude oil markets has not fully subsided. Although previously built hedging positions were partially unwound during this rally, with the VIX briefly retreating near 17, cross-asset volatility remains relatively elevated.

This means whether the "breathing room" brought by CoreWeave can evolve into a trend-based recovery depends on whether subsequent earnings, the funding market, and the macro interest rate environment can simultaneously deliver positive answers.

Chart: Intraday performance of US 10-year Treasury futures and 10-year Treasury yields, with yields nearing 5% during trading hours.

Overall, CoreWeave's earnings temporarily restored market confidence in AI compute demand and financing capabilities, creating a more favorable sentiment window for Anthropic's IPO and Nvidia's earnings report. However, transitioning from "short-term short covering" to "long-term fundamental repricing" still requires answering four critical questions: funding, supply, GPU residual values, and interest rates.

Anthropic and Nvidia next week will serve as two critical tests to determine whether this AI rally can continue.