Back to News
Market Impact: 0.35

Down 50% From Its All-Time High, Should You Buy CoreWeave Before 2025 Is Over?

CRWVNVDAMSFTMETANFLXNDAQ
Artificial IntelligenceTechnology & InnovationCompany FundamentalsCorporate EarningsIPOs & SPACsAnalyst InsightsInvestor Sentiment & PositioningCorporate Guidance & Outlook
Down 50% From Its All-Time High, Should You Buy CoreWeave Before 2025 Is Over?

CoreWeave, the AI-focused cloud provider that rents Nvidia GPUs to customers including OpenAI, Microsoft and Meta, has rallied since its March 2025 IPO but remains about 50% off its June peak; the company is pursuing aggressive capacity builds (Q1–Q3 capex of $1.9B, $2.9B and $1.9B respectively) while generating only $1.4B in Q3 revenue and burning over $8B in free cash flow over the past 12 months. Its business benefits from a large revenue backlog ($55.6B, with 40% due in the next 24 months), but structural headwinds — GPUs’ short 1–3 year lifespan, rapid obsolescence from new Nvidia launches and therefore continuing high replacement capex — mean CoreWeave will need to prove profitability soon; the author doubts it will break even in 2026 and recommends waiting for clearer progress toward sustainable profits.

Analysis

CoreWeave (CRWV) is an AI-focused cloud provider that rents Nvidia GPUs to large generative-AI customers including OpenAI, Microsoft and Meta; the stock has rallied 125% since its March 2025 IPO and peaked up ~360% in June but currently trades roughly 50% below that high, reflecting volatile sentiment. The company is pursuing an aggressive buildout: capital expenditures were $1.9 billion in Q1, $2.9 billion in Q2 and $1.9 billion in Q3, while Q3 revenue totaled $1.4 billion, and over the past 12 months CoreWeave spent roughly double what it earned in revenue and burned over $8 billion in free cash flow. CoreWeave reports a $55.6 billion revenue backlog with 40% of that backlog due within 24 months, which supports forward revenue visibility and underpins the bullish case if backlog converts at healthy margins; the author flags the potential for 2026 to be the inflection year but explicitly doubts a 2026 profitability turnaround. Structural risks are material: GPU lifespans are 1–3 years under heavy AI workloads and rapid Nvidia product cycles drive near-perpetual replacement capex, meaning capital intensity and operating losses could persist absent obvious margin expansion or FCF stabilization.

AllMind AI Terminal