CoreWeave raised its 2026 revenue and adjusted operating-income outlook after third-quarter revenue climbed 168% from a year earlier, strengthening the company’s case that demand for AI computing capacity is arriving faster than its earlier plans assumed. The upgrade also lifts the lower end of a capital-spending range that now calls for as much as $39 billion this year, putting fresh attention on how the cloud-infrastructure provider will finance its buildout.
For the quarter ended Sept. 30, CoreWeave reported revenue of $3.661 billion, compared with $1.365 billion a year earlier. Its new full-year revenue forecast is $13 billion to $13.3 billion, up from $12.4 billion to $13.2 billion previously, according to the company’s third-quarter earnings release. The results exceeded the $3.55 billion revenue consensus compiled by LSEG, as reported by CNBC.
The figures offer a more detailed test of the economics behind the AI infrastructure boom. CoreWeave is converting committed customer demand into sharply higher reported sales, but it is doing so through an unusually capital-intensive expansion whose funding costs remain substantial.
Guidance rises alongside deployment plans
CoreWeave lifted its 2026 adjusted operating-income forecast to $1.275 billion to $1.375 billion from a prior range of $960 million to $1.15 billion. The company said the midpoint of the new range was 26% above the midpoint of its earlier projection. It also increased the low end of its capital-expenditure outlook to $37 billion from $35 billion, leaving the top end unchanged at $39 billion.
Management attributed the higher spending plan to an expectation that it can deploy and deliver infrastructure sooner than anticipated. That means CoreWeave’s investment program is running at a scale far above current annual revenue: the $38 billion midpoint of the capital-expenditure range is roughly 2.9 times the $13.15 billion midpoint of the new revenue outlook. The comparison does not measure cash spending in the same period or profitability, but it illustrates the amount of physical capacity the company is seeking to bring online.
At quarter-end, CoreWeave said it had more than 1.7 gigawatts of active power and still expected to exceed 1.85 gigawatts by year-end. Power availability is a critical operating constraint for AI data centers, where dense clusters of graphics processors require both electricity and facilities capable of handling the associated cooling and networking loads.
The company reported approximately $132.5 billion in revenue backlog as of Sept. 30. That measure includes remaining performance obligations and other estimated future revenue from committed customer contracts. It is a sizable demand indicator, but not a booked-revenue equivalent: CoreWeave says the amount remains subject to delivery and service-availability requirements. Investors assessing the figure will need to distinguish contracted demand from the execution required to energize data centers, install equipment and begin delivering service.
Growth has not eliminated the financing question
CoreWeave posted GAAP operating income of $207 million, up from $52 million a year earlier, and adjusted EBITDA of $2.256 billion, or a 62% adjusted EBITDA margin. Yet it recorded a GAAP net loss of $709 million, wider than the $110 million loss a year earlier. The company reported $846 million of net interest expense during the quarter, while a Reuters report put CoreWeave’s debt at roughly $46.9 billion at quarter-end.
The gap between operating measures and net income is especially important for a company using large amounts of debt and other financing to acquire computing equipment and construct data-center capacity. Adjusted EBITDA and adjusted operating income exclude or alter items under the company’s non-GAAP methodology, while interest costs still affect GAAP net income and cash demands. CoreWeave said it could not provide a forward-looking reconciliation of adjusted operating-income guidance to the most directly comparable GAAP measure because of uncertainty around reconciling items.
Per-share comparisons similarly require care. CoreWeave’s release showed a GAAP loss of $1.27 per basic and diluted share. CNBC, citing LSEG data, reported an adjusted loss of 95 cents a share against an expected adjusted loss of $1.16. Those are different accounting measures, so the 95-cent adjusted result should not be read as a revision of the company’s $1.27 GAAP loss.
During the quarter, CoreWeave raised about $4.2 billion through convertible senior notes and closed a $2.6 billion financing facility. On the earnings call, Chief Executive Michael Intrator said the company expected about $9 billion of customer prepayments and deposits over the next three months. He also said CoreWeave expected customer prepayments and hardware-partner financing to provide most of the capital it needs in 2027, with a limited amount of additional fundraising, according to CNBC.
Finance chief Nitin Agrawal said on the same call that CoreWeave did not intend to issue additional parent-level unsecured debt through 2027 at current yields and projected less than $5 billion in parent-level debt financing through then, CNBC reported. The comments signal that management sees customer-linked and asset-backed funding as preferable to more expensive unsecured borrowing while it expands capacity.
Market reaction reflected separate trading windows
The stock response was sharp, though reports describing it covered different periods. Reuters said shares rose 5% in extended trading after the Nov. 10 results. The following day, Yahoo Finance reported that the shares were up as much as 25% during regular trading. The two percentages should not be combined; they describe after-hours and intraday moves, respectively.
CoreWeave has said it expects to become profitable on an adjusted basis in the first half of 2028, Reuters reported. Until then, the company’s higher forecast, backlog conversion and ability to fund faster deployments without adding costly parent-level debt will remain the practical measures investors use to judge whether its growth can support the scale of its infrastructure commitments.
