CoreWeave built its entire business on a bet that demand for GPU cloud capacity would outrun every other consideration, including profitability. So far, the bet is half right: the demand is real. The profitability isn’t.
Shares trade at $90.30, a market cap near $49.3 billion, against a trailing EPS of -$2.02 and a negative P/E of -44.75x. The company’s free cash flow position is arguably the more revealing number: price-to-free-cash-flow sits at -4.96x, meaning CoreWeave is burning cash relative to its market value, not generating it — standard for a company still in the land-grab phase of infrastructure buildout, but a meaningfully different risk profile than a name generating positive free cash flow at scale.
The balance sheet backs up the “growth at any cost” read. Debt-to-equity sits at 6.79x, extraordinarily leveraged even by capital-intensive infrastructure standards — this is a company financing GPU clusters with debt against long-term contracts, not equity or retained earnings. Gross margin, at 71.7%, actually looks healthy on the surface, which tells you the losses aren’t coming from the core unit economics of renting out compute — they’re coming from the depreciation, financing costs, and buildout expense layered on top as the company races to add capacity ahead of demand it’s betting will materialize.
Price-to-sales of 7.0x puts CoreWeave in the same valuation neighborhood as established, profitable cloud infrastructure names — a multiple that assumes today’s unprofitable land-grab converts into tomorrow’s high-margin utilization business, the way AWS did over a decade-plus timeline, just compressed into a fraction of the time.
Analyst sentiment reflects genuine uncertainty rather than conviction in either direction. The consensus rating is buy, but 36.4% of covering analysts sit at neutral — one of the highest neutral shares in this entire group of names — with 4.5% outright bearish. The average price target of $139.18 implies upside of roughly 54.1%, but that number is doing a lot of work to average out analysts who see this as the picks-and-shovels AI infrastructure trade against analysts who see a company whose capital structure can’t survive a demand air pocket. With debt-to-equity near 6.8x, there isn’t much room for the second group to be even partially right before the equity feels it first.
