With CoreWeave fueling the AI frenzy, can Nebius (NBIS.US) defend the AI computing power rally with explosive performance?

According to Zhitong Finance APP, CoreWeave (CRWV.US), one of the leaders in NeoCloud, has significantly alleviated market anxiety and skepticism about the new cloud business model with robust performance growth, computing power orders much more optimistic than market expectations, and financing arrangements. The company’s strong financial results and explosive orders have driven a broad rebound in AI computing power industrial chain stocks across the U.S. stock market.

Another new cloud leader, Nebius Group NV (NBIS.US), a European tech frontrunner that has recently shifted to focus on providing cloud-based AI computing resources, is now under pressure to prove that computing power price increases can be fulfilled in tandem with better-than-expected capacity delivery and faster-than-expected capital inflow, especially after the NVIDIA CEO Jensen Huang was hit by another negative incident causing NVIDIA’s share price to plunge after hours and dragging the entire AI computing power sector down. Nebius Group will release its earnings report and forward guidance before U.S. market open on November 12, Beijing time.

By the close of U.S. stock markets on Wednesday, CoreWeave's strong financial results and financing plans led to a rally in storage chips, AI chips, optical interconnects, and new cloud stocks; CoreWeave and Nebius rose about 24.7% and 15.4% respectively. However, after hours, media reports about another extramarital affair involving Jensen Huang and rumors of NVIDIA’s board discussing his departure caused NVIDIA to drop more than 8%, with most AI computing-related stocks coming under pressure as well.

At the same time, CoreWeave’s explosive performance, Anthropic’s IPO prospectus revealing $16.956 billion in quarterly revenue, and especially the AI application leader’s multi-year, $900.9 billion procurement and leasing commitments continue to support expectations for a surge in computing power demand. The next question for Nebius to answer is whether robust demand can continue to be converted into deliverable, billable, and return-generating computing power.

Nebius focuses on integrating GPUs, high-speed optical interconnects, storage, and scheduling software into training and inference services. With its business structure based primarily on 1—3 year contracts, Nebius aims to capitalize more quickly on price increases in computing power. Truist maintains a "Buy" rating and a $355 target price, optimistic about this repricing capability; Citi has raised its forward revenue forecasts but lowered the target price, reflecting the market's higher demands for financing and valuation. If this earnings report simultaneously delivers better-than-expected revenue, faster capacity ramp-up, and clear follow-up financing arrangements, Nebius may extend the counter-offensive initiated by CoreWeave and further validate the profitability of new cloud, once again making fundamentals the core bullish pricing basis for AI computing power investments.

CoreWeave Ignites the Computing Power Rally, Nebius Poised to Test Growth Strength

As of U.S. market close on November 11 (EST), CoreWeave rose 24.7%, Nebius rose 15.44%, closing at $154.69; the Nasdaq 100 and Philadelphia Semiconductor Index climbed 2.28% and 5.66% respectively. Storage giant Micron and Sandisk rose 11.89% and 13.69%, leading optical interconnect firms Lumentum and Coherent rose 10.67% and 9.25%, and capital rotated back into core infrastructure links of AI computing power including AI chips, storage, high-speed optical connections, computing power leasing, CPUs, PCBs, and advanced 3D packaging.

The key support for this rally is that computing power demand is now being validated doubly by revenue and funding sources. CoreWeave's Q3 revenues reached $3.661 billion, up 168% year-on-year, with full-year guidance raised to $13–13.3 billion, and the company expects to receive about $9 billion in customer prepayments and deposits over the next three months. Management also specified that at current yield levels, construction will mainly be supported by Delayed Draw Term Loans (DDTL) aligned with specific customer contracts, reducing reliance on high-cost unsecured financing. The investment enthusiasm signaled here is clear: major customers are committing to helping computing power suppliers raise construction funds, lowering the risk of order expansion being hindered by financing bottlenecks.

The AI application side has also provided new demand cues. Anthropic’s prospectus reported Q3 revenues of $16.956 billion, up about 1127% year-on-year and 46.7% quarter-on-quarter; procurement and leasing commitments total $900.9 billion, most of which are for computing power and infrastructure, including $451.6 billion expected by the end of 2030 and the rest mainly from 2031—2045. This highlights the link between model commercialization and long-term computing power procurement, though realization of these commitments as supplier revenue still depends on specific delivery arrangements. As such, Anthropic’s IPO tests not only application-layer valuations but also the capital market’s willingness to support ongoing investments in computing power.

However, it’s worth noting that a new trading variable emerged after Wednesday’s market close. Media reports early Thursday Beijing time claimed Jensen Huang was embroiled in another extramarital affair, with reports that NVIDIA’s board was discussing his future; as a result, the stock fell more than 8% after hours, and Nebius dropped about 4.1%. Should Nebius' earnings surprise on revenue, delivery, and financing, it could greatly boost market confidence in computing power demand after CoreWeave's strong performance, and may even help attenuate negative sentiment tied to NVIDIA management uncertainty.

Nebius, much like CoreWeave, holds core value in the AI computing power supply chain by organizing GPUs, electricity, high-speed networking, storage, and scheduling software into reliably deliverable AI cloud services. Training large models requires massive GPU synchronization of parameters and data, while inference and agents continually consume compute, VRAM, and storage resources—buying chips alone cannot solve cluster efficiency and reliability issues. Nebius reduces customer deployment difficulty through InfiniBand interconnects, training schedulers, and managed inference platforms, addressing capacity demand clients can't build themselves in time. In tight supply, rapid delivery and shorter contract terms create repricing opportunities, enabling revenue per megawatt to respond faster to demand growth.

Nebius Earnings Preview: Focus Shifts to Computing Power Demand, Contracts, and Financing Plans

Nebius Group will announce its Q3 results before the U.S. market opens Thursday (Beijing time). Wall Street analysts unanimously expect a loss per share of $0.63 and revenues of $907.17 million for the quarter.

Last quarter, the Netherlands-based AI infrastructure provider posted better-than-expected second-quarter results and reiterated its 2026 full-year targets: $7–9 billion in annualized revenue run rate, $3–3.4 billion in group revenues (midpoint $3.2 billion), and about 40% group adjusted EBITDA margin. Consensus expectation for full-year revenue is $3.35 billion.

Veteran Wall Street analysts remain bullish, focusing on Nebius’s future build-out, financing plans, and the pricing of new and shorter-term contracts.

Truist Securities believes Nebius is particularly well-positioned to capture rising computing power prices, as its business is mainly structured around 1- to 3-year contracts. The firm expects these shorter contracts to give the company greater flexibility to profit from price hikes.

This well-known Wall Street broker projects Nebius’s annualized revenue run rate will exceed $4.8 billion this quarter, with installed capacity reaching 730MW and operating capacity about 600MW. It reiterates a "Buy" rating and $355 target price.

Citi’s Equity Research division has also raised its 2027–2030 revenue estimates for Nebius by 20%, and expects annualized revenue run rate per megawatt to accelerate in coming years amid sustained computing power demand growth. However, despite maintaining a "Buy" rating, Citi lowered its target price from $324 to $265 per share.

Overall, Seeking Alpha’s Quant Rating, its in-house analysts, and Tipranks-compiled Wall Street ratings all remain bullish, unanimously granting the stock the equivalent of a “Buy”—the most optimistic and positive rating.

Compared with industry peer CoreWeave, Nebius carries much lower overall credit liabilities and operating leverage, so its share price has proven much more resilient, especially as CoreWeave’s share price has nearly halved. Nebius’s share price is up over 82% so far this year, handily outperforming the S&P 500 benchmark’s nearly 10% gain in the same period.

However, Seeking Alpha’s analyst Lee Bailey cautions that Nebius faces high capital expenditure, equity dilution, and equipment upgrade risks, assigning it a “Sell” rating. Bailey adds that while Nebius’ major contracts with Microsoft and Meta provide rare revenue visibility, rapid growth also exposes the company to risks around execution, delivery, and potential funding breaks linked to AI infrastructure financing.

Bailey concludes, “The current pricing and margin profile are excellent, but their sustainability is uncertain as hardware upgrade cycles and future competition could erode returns.”

Over the past three months, EPS expectations have seen four upward and three downward revisions, while revenue forecasts have undergone seven upward and five downward revisions.

Computing Power Price Increases Expand Revenue, but Capacity Expansion and AI Infrastructure Financing Costs Drive Shareholder Returns and AI Power Rally Trajectory

Nebius plans to announce its Q3 results before the U.S. market opens on November 12. According to consensus from Seeking Alpha, expectations are for $907.17 million in revenue and a $0.63 per share loss—representing about 521% year-over-year growth and 55.8% sequential growth. The market’s real focus is whether the new capacity is already contributing to revenue, and whether the company can fulfill its year-end $7–9 billion annualized revenue run rate (ARR) target.

It’s important to note that ARR here is calculated as the last month’s AI cloud revenue in the quarter multiplied by 12, reflecting the company’s business scale at the end of the period. Truist expects third-quarter ARR to exceed $4.8 billion, equivalent to 730 MW of energized capacity and approximately 600 MW of operating capacity—the gap between the two being the key to converting capacity into billable services.

The consensus among bullish institutions is based on constraints on computing power supply and frequent repricing for short-term contracts. Truist maintains a “Buy” and $355 target price, optimistic that a business structure centered on 1- to 3-year contracts can capture price hikes; Citi raised its 2027–2030 revenue estimates by 20% but lowered the target price from $324 to $265, reiterating “Buy”; and Bank of America recently reiterated “Buy” and a $310 target price, bullish on capacity expansion and higher prices for shorter contracts.

Based on the November 11 close of $154.69, these three targets imply potential upside of roughly 129.5%, 71.3%, and 100.4% respectively. Citi’s combination of raising revenue forecasts but lowering its target price especially shows that operational growth and equity valuation are now being driven by different factors; strong demand will still require financing and capital returns to support it.

This divergence can to some extent be explained by financing structure. In October, Nebius announced $3.5 billion in senior secured financing from Citi, maturing in 364 days, with an initial interest rate of term SOFR plus 275 basis points, increasing according to agreement. So, while short-term construction funds are replenished, the market still needs to assess how long-term financing will be maintained. For GPU cloud companies, rental rate increases boost revenues, but equipment utilization determines the extent to which fixed costs are diluted, while depreciation, renewal expenses, and interest ultimately determine net returns. The most valuable earnings surprise this time would be a combination of stronger AI computing power order data, higher billable capacity, better contract pricing, and financing arrangements that better match the delivery pace.