CoreWeave, Inc. 10-Qquarter ended 2026-03-31

Executive Summary

CoreWeave sells specialized cloud computing capacity to companies developing and running artificial intelligence, earning revenue mainly through committed service contracts. It is expanding its NVIDIA-based infrastructure to serve large customers such as OpenAI and Meta, with $98.8B of remaining performance obligations supporting the buildout. Q1 revenue more than doubled to $2.08B, but rising infrastructure costs and interest expense widened losses. The opportunity is to convert contracted demand into productive capacity; the threat is committing capital faster than facilities, customers and financing can generate sustainable returns. For investors, the decisive question is whether delivery and eventual profitability justify the growing funding burden.

What's Happening in the Business

• Customer expansion: Existing customers supplied 38% of the revenue increase, with new customers providing the remainder. Committed services generated 98% of revenue, making delivery against contracts—not simply winning orders—the immediate growth driver.

• Capacity buildout: Net property and equipment rose to $36.4B from $30.6B at year-end. Higher rent, power costs and depreciation accompanied new capacity, showing that rapid deployment is currently compressing margins rather than producing operating leverage.

• Longer-term delivery commitments: Management expects 36% of remaining performance obligations to become revenue through March 2028. That creates visibility, but recognition estimates incorporate potential customer credits, delivery delays and capacity resale; contracted demand is not equivalent to immediate revenue or cash.

Key Financial Highlights

• Cash conversion: Operating cash flow benefited from $1.04B of accounts-receivable inflows and $960M from accounts payable and accrued expenses. Capital expenditures, including capitalized internal-use software, reached $7.70B versus $1.41B; management cautioned that operating-cash-flow growth partly reflected timing.

• Equity funding and dilution: January’s NVIDIA placement generated $2.0B gross proceeds. Shares outstanding increased 6% to 532M during the quarter, while stock-based compensation, net of capitalized costs, was $153M. April brought another $1.0B gross equity issuance, $4.0B principal of convertible notes and $2.75B principal of senior notes; the company also repaid its $1.5B revolver balance.

• Liquidity composition: Reported liquidity reached $11.1B, but $8.83B consisted of undrawn revolving and delayed-draw facilities rather than cash. Continued investment therefore depends substantially on access to borrowing.

Risk Assessment

• Financial-reporting controls: Disclosure controls were ineffective because material weaknesses in technology controls, segregation of duties and staffing remained unresolved; no material misstatement had resulted, but remediation still requires implementation and testing.

• Funding and refinancing: Scheduled principal payments of $6.07B for the remainder of 2026 exceed quarter-end cash and cash equivalents of $2.24B, making financing access critical alongside the continuing infrastructure cash deficit.

• Capacity execution and fixed commitments: Delivery delays were announced in November 2025, while executed leases not yet commenced carry $40.7B of estimated undiscounted payments; further power, construction or commissioning delays could defer revenue without removing commitments.

• Customer concentration: Two customers generated 65% of Q1 revenue, so nonperformance, lower purchasing or missed delivery requirements could leave expensive capacity unused.

• Hardware dependence and competition: All deployed graphics processors are NVIDIA products, while larger competitors including Amazon Web Services, Google Cloud and Microsoft Azure can pressure pricing or build competing capacity; supply disruption or weaker AI spending could undermine returns on installed equipment.

What to Watch

• Capacity delivery: The infrastructure asset base is expanding rapidly after previously announced delays. Later filings should show timely commissioning and contract revenue recognition; further delays, credits or idle capacity would weaken the investment case.

• Customer breadth: New customers supplied most incremental revenue, but two customers still dominate sales. Broader customer contributions would reduce dependence; renewed concentration or customer retrenchment would increase exposure to individual contracts.

• NVIDIA Rubin deployment: Management expects deployment in the second half of 2026. On-time availability and customer adoption would support CoreWeave’s technology positioning; postponement or weak uptake would challenge the spending required to keep its platform competitive.

• Funding versus self-financing: April’s debt and equity raises extended funding after a deeply negative free-cash-flow quarter. Improving cash generation relative to capital expenditures would reduce dependence on external capital; persistent deficits would imply further leverage or dilution.