
CoreWeave's stock has fallen 52% from its 52-week high as the cost of funding its rapid growth—driven by heavy borrowing for data centers—outpaces revenue gains. While Q1 revenue jumped 112% year-over-year to $2.1 billion(約3400億円), interest expense more than doubled to $536 million(約860億円) and net losses widened to $740 million(約1200億円). The challenge is compounded by Meta Platforms planning a competing cloud business to sell AI computing capacity, even as Meta remains one of CoreWeave's largest customers. The company has a record $99.4 billion(約16兆円) revenue backlog, but investors need to see that backlog convert into profit while debt costs stabilize.
Summaries like this, in your inbox every morning.
Sign up free →What happened
CoreWeave, an AI cloud provider, has seen its stock fall from a 52-week high of $153.20 to $72.91 as of Thursday. Revenue surged 112% year-over-year to $2.1 billion(約3400億円) in Q1, but interest expense more than doubled to $536 million(約860億円) (from $264 million(約420億円) in the prior year), and net loss widened to $740 million(約1200億円) from $315 million(約500億円).
Why it matters
The company borrows heavily to build data centers, and that debt service is growing about as fast as the business itself—a dynamic that is pressuring the stock. Additionally, Meta Platforms is planning an internal cloud business called Meta Compute to sell surplus AI computing capacity, which directly competes with CoreWeave's core offering, even though Meta is one of CoreWeave's largest customers under an agreement worth about $21 billion(約3.4兆円) through 2032.
What to watch
CoreWeave's revenue backlog reached $99.4 billion(約16兆円) as of March 31 (the strongest bookings quarter in company history), and management believes active power will exceed 8 gigawatts by 2030 (up from 1 gigawatt in Q1). Investors are waiting for evidence that this backlog converts into revenue at healthy margins and that interest costs grow far more slowly, proving the debt-heavy model can scale toward profitability.
CoreWeave closed Thursday at $72.91, down 52% from its 52-week high of $153.20, as the AI cloud infrastructure provider faces mounting pressure from the cost of its own growth. The company's Q1 results, reported in May, laid bare the tension: revenue surged 112% year-over-year to $2.1 billion(約3400億円), but interest expense more than doubled to $536 million(約860億円), up from $264 million(約420億円) in the year-ago quarter. Net loss widened sharply to $740 million(約1200億円) from $315 million(約500億円). Those results triggered a 10% stock decline as investors reacted to disappointing revenue guidance and expanding spending forecasts.
The underlying issue is structural. CoreWeave borrows heavily to construct data centers that support its AI computing rental business, and the debt service bill is growing about as fast as the business itself. This week brought fresh selling pressure: shares fell 3.5% on Wednesday and again on Thursday as AI infrastructure stocks sold off broadly. CEO Michael Intrator's insider sales—approximately 369,000 shares for roughly $31 million(約50億円) in early July and about 308,000 more for roughly $25 million(約40億円) on July 14—added to negative sentiment, though the sales were executed under a prearranged trading plan adopted the prior year.
Complicating the outlook is Meta Platforms' entry into the same market. Bloomberg reported on July 1 that Meta is planning an internal cloud business called Meta Compute to sell surplus AI computing capacity to enterprise customers—directly competing with CoreWeave's core offering. The timing is awkward: Meta is also one of CoreWeave's largest customers, having expanded their relationship in April with an agreement worth about $21 billion(約3.4兆円) through 2032. That contract once appeared to be a significant validation; now it raises the risk that CoreWeave's biggest customer could become a major competitor.
Demand itself is not the problem. CoreWeave's revenue backlog reached $99.4 billion(約16兆円) as of March 31, which management characterized as the strongest bookings quarter in company history. Active power usage topped 1 gigawatt in Q1, and the company projects it will exceed 8 gigawatts by 2030. The market's concern is whether that enormous backlog will convert to revenue at healthy margins, and whether interest costs can grow far more slowly than they have been, allowing the debt-heavy model to scale toward profitability. Until CoreWeave demonstrates both, the pattern of the past month is likely to persist: strong demand headlines followed by reminders of the high cost of funding that growth.
CoreWeave is caught in a bind that reveals a structural tension in the high-growth, capital-intensive AI infrastructure business. The company's revenue growth (112% year-over-year to $2.1 billion(約3400億円) in Q1) would normally be celebrated, and its $99.4 billion(約16兆円) backlog shows genuine customer demand. Yet the financing model underlying this growth is becoming a liability: interest expense more than doubled to $536 million(約860億円) while net losses expanded to $740 million(約1200億円). The market is punishing CoreWeave not for lack of demand but for the cost of serving it.
The Meta Compute announcement on July 1 crystalized investor anxiety. Meta is simultaneously CoreWeave's largest customer (under a $21 billion(約3.4兆円) commitment through 2032) and now a future competitor. This threatens the margin story—even if CoreWeave converts its backlog, it may do so against a customer-turned-rival that has vastly deeper pockets and can amortize AI infrastructure across its own business. The stock's 52% decline from $153.20 reflects the market's bet that CoreWeave will struggle to prove its debt-heavy model can reach sustainable profitability before cash costs overwhelm the top-line story.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
No discussion yet for this article
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack