CoreWeave surged 119% over the past 12 months, outperforming its industry peers as enterprises raced to secure high-performance GPU capacity for generative AI workloads. The company reported a tripling of revenue in fiscal 2024, driven by a $55 billion sales backlog in Q3, up 100% year-over-year. Management expects to reach 850 megawatts of active power capacity by the end of 2025, positioning CoreWeave to capitalize on a multi-trillion-dollar AI infrastructure market projected through 2030.

CoreWeave’s rapid expansion requires substantial capital; net capital expenditures exceeded $7 billion in the past 12 months, contributing to a total debt load of approximately $12 billion. The company reported net losses of $863 million in 2024 and operating losses of $748 million over the first three quarters of 2025. Analysts warn that sustained high cash burn—estimated at $19 billion through 2026—could strain liquidity if GPU depreciation accelerates or if demand growth slows.

Key agreements underpin CoreWeave’s business model: Nvidia invested $300 million pre-IPO for a 7% equity stake and committed to purchase any unsold capacity under a $6.3 billion contract running through 2032. This arrangement secures revenue stability, while large clients such as OpenAI and Microsoft contribute to a diversified customer base. Goldman Sachs initiated coverage with a Neutral rating, citing an architectural edge, and Moody’s forecasts $3 trillion in global data-center investment over five years. Execution risks remain around power build-out delays and debt servicing, but valuation metrics—trading at under eight times trailing sales—suggest potential upside if CoreWeave can demonstrate sustainable profitability.