CoreWeave and Nebius have delivered one of the clearest recent signals that demand for AI computing capacity remains strong.
The two companies sit in the fast-growing NeoCloud segment: GPU-heavy cloud infrastructure providers built around AI training and inference workloads. Their latest earnings show not only rapid revenue growth, but also large customer commitments, rising capacity requirements and continued infrastructure investment.
For investors, the key question is shifting from whether AI demand exists to whether NeoCloud providers can build enough capacity—and do so profitably—to meet it.
CoreWeave Q2 Earnings: Backlog Reaches $104 Billion
CoreWeave reported second-quarter revenue of $2.58 billion, up about 112% year over year from $1.21 billion.
Adjusted EBITDA reached approximately $1.51 billion, although the company remained unprofitable on a GAAP basis, reporting a $626 million net loss. Net interest expense climbed to roughly $640 million as CoreWeave continued financing its aggressive infrastructure expansion.
The stronger signal came from contracted demand.
According to CoreWeave's
official Q2 2026 earnings release, revenue backlog reached approximately
$104 billion at the end of June. The company also said this figure excluded more than
$25 billion in net new customer commitments added in early Q3.
CoreWeave continued expanding physical capacity as well. Active power increased by nearly 500 MW during the quarter to around 1.5 GW, while total contracted power reached approximately 3.7 GW.
These figures suggest CoreWeave's immediate problem is not finding customers. It is building infrastructure quickly enough to serve demand already under contract.
CoreWeave Is Spending Heavily to Keep Up
Strong demand comes with an equally large capital requirement.
CoreWeave now expects
2026 capital expenditures of $35 billion to $39 billion, alongside full-year revenue guidance of
$12.4 billion to $13.2 billion. Its Q3 revenue outlook stands at $3.45 billion to $3.60 billion, according to the company's
latest earnings outlook.
That spending highlights the central tension in the NeoCloud model.
Large AI customers want more GPU capacity, but meeting that demand requires enormous upfront spending on accelerators, data centers, power infrastructure and financing.
CoreWeave therefore has to prove that rapid capacity expansion can eventually produce sustainable margins and cash flow—not just higher revenue.
Nebius Q2 Earnings: AI Cloud Revenue Jumps 514%
Nebius delivered even faster growth from a smaller base.
Q2 group revenue reached $582.3 million, up 454% year over year and 46% sequentially. Its core Nebius AI Cloud business generated approximately $574.9 million, representing about 98% of group revenue and rising 514% year over year.
Annualized run-rate revenue reached approximately $3.0 billion by the end of June.
The company signed four major AI Cloud deals during Q2, each with an average total contract value above $1 billion. Total contract value from deals signed during the quarter increased nearly fourfold from Q1, while contract value from new customers rose more than ninefold.
Those figures provide another important signal: AI infrastructure demand is translating into large, long-term customer commitments rather than just short-term GPU usage.
Customers Are Paying Upfront for AI Capacity
Nebius disclosed an even stronger indicator of tight supply.
Around 70% of deals signed during Q2 included customer prepayments, and the company expects to receive more than $9 billion in customer prepayments during 2026.
Total customer commitments now exceed $40 billion.
Nebius also said current demand is strong enough that it could effectively sell its planned 2027 capacity today under current commercial terms. Instead, the company is keeping some future capacity available for customers that may require compute on shorter notice.
To support that demand, Nebius raised its year-end 2026 contracted power target to 5 GW and plans to deploy more than 1 GW of additional capacity per year from 2027.
This is significant because customers are not simply expressing interest in AI compute. They are signing multi-year contracts, reserving future capacity and providing cash before that capacity is fully available.
What CoreWeave and Nebius Tell Us About AI Demand
For much of the AI boom, Nvidia's data-center revenue was the clearest proxy for underlying demand.
CoreWeave and Nebius now provide another layer of evidence.
The infrastructure chain increasingly looks like this:
AI models → GPUs → AI cloud capacity → networking → optical interconnect → power and cooling
CoreWeave and Nebius sit directly between GPU suppliers and companies that need large amounts of compute for AI training and inference.
Their latest results suggest that demand at this layer has not yet reached an obvious ceiling.
CoreWeave has more than $100 billion in backlog. Nebius is growing AI Cloud revenue above 500%, signing billion-dollar contracts and collecting substantial customer prepayments.
Together, those figures suggest the AI compute market remains capacity constrained rather than demand constrained.
That distinction matters for the broader AI infrastructure trade.
If GPU cloud providers continue adding capacity aggressively, spending can keep flowing downstream into servers, networking, optical components, power infrastructure and cooling systems.
Is the NeoCloud Boom Still Accelerating?
For now, the operating data point in that direction.
Several indicators are strengthening at the same time:
Revenue is rising rapidly. Contract sizes are getting larger. Customers are reserving future capacity. Power targets are moving higher. And both companies continue committing billions of dollars to infrastructure expansion.
That is stronger evidence than management simply saying that “AI demand remains strong.”
It shows that customers are willing to commit capital to secure access to compute.
The market reaction reflected that optimism. CoreWeave shares rose sharply following its results, while Nebius also rallied after reporting stronger-than-expected growth and demand. Broader AI infrastructure names moved higher as investors interpreted the earnings as another sign that AI spending remains resilient.
The Bigger Risk Is Execution, Not Demand
Strong demand does not make NeoCloud stocks low risk.
CoreWeave plans to spend as much as $39 billion this year, while Nebius continues investing heavily in GPUs, data centers and power capacity.
That creates three major challenges.
First, signed contracts only become valuable if companies can deliver the infrastructure required to recognize the revenue.
Second, financing matters. Rapid expansion can create heavy interest expense, depreciation and cash burn even when headline revenue growth remains strong.
Third, today's shortage could eventually become excess capacity if AI spending slows, GPU supply improves faster than expected or large customers increasingly build their own infrastructure.
Backlog also should not be treated as the same thing as current revenue. Contracts still depend on capacity delivery, service availability and customer execution.
The next phase of the NeoCloud trade therefore becomes much more fundamental.
Investors need to watch whether backlog becomes revenue, revenue becomes margin, and margin eventually becomes cash flow.
What Should Investors Watch Next?
Future CoreWeave and Nebius earnings should be judged on more than headline revenue growth.
The most important indicators are:
Revenue growth relative to new capacity, contracted and active power, customer commitments, utilization, pricing and gross margins.
Financing costs and capital expenditures will matter just as much.
If utilization and revenue grow faster than depreciation, infrastructure costs and interest expense, NeoCloud providers can begin to demonstrate real operating leverage.
If spending continues accelerating while margins or utilization weaken, investors may become less willing to reward growth alone.
That is increasingly the dividing line across the AI infrastructure market.
The question is no longer simply which companies have AI exposure.
It is which companies can convert that exposure into orders, backlog, revenue and eventually cash flow.
Explore CoreWeave, Nebius and Stock-Related Markets on MEXC
Investors following the AI infrastructure theme can use MEXC to monitor related U.S. equity markets. The
MEXC US Stocks market hub provides access to market information for listed companies, including a dedicated
NBIS stock page.
Availability and product access may vary by region.
FAQ
When did CoreWeave report Q2 2026 earnings?
CoreWeave reported its Q2 2026 results on August 11. Revenue reached approximately $2.58 billion, while revenue backlog stood at about $104 billion.
When did Nebius report Q2 2026 earnings?
Nebius reported Q2 2026 results on August 12. Group revenue reached $582.3 million, up 454% year over year, while Nebius AI Cloud revenue increased 514%.
Why are CoreWeave and Nebius important to the AI trade?
Both companies provide GPU-intensive cloud infrastructure for AI workloads. Their revenue growth, contracts and capacity expansion offer a direct read on demand for AI computing beyond traditional hyperscalers.
What is a NeoCloud?
A NeoCloud is a cloud infrastructure provider focused heavily on high-performance computing and GPU-based AI workloads. CoreWeave and Nebius are two of the most closely watched publicly traded companies in this segment.
Is AI cloud demand still growing?
The latest results from CoreWeave and Nebius suggest demand remains strong. However, the next test is whether growing commitments can translate into sustainable revenue, margins and cash flow.