CoreWeave (NASDAQ:CRWV) has been a quintessential AI growth story, delivering specialized GPU cloud infrastructure that powered the sector’s insatiable compute needs. Yet mounting worries about a ballooning debt load, persistent questions surrounding AI’s long-term sustainability, and fears of circular financing — where suppliers like Nvidia (NASDAQ:NVDA | NVDA Price Prediction) prop up demand for their own chips — quickly weighed on sentiment.
After hitting an all-time high of $187 per share last June, shares have plunged more than 56% and now trade near $81. Despite the overhang, CoreWeave keeps delivering one strategic victory after another, transforming what looks like a battered stock into a compelling bargain for investors who can look past the noise.
A $2 Billion Equity Infusion Deepens a Critical Partnership
In late January, Nvidia poured another $2 billion into CoreWeave Class A shares at $87.20 each, building on earlier stakes from 2023 and the 2025 IPO. This cash directly bolsters the balance sheet and accelerates CoreWeave’s aggressive rollout of data center capacity, land acquisitions, and power infrastructure.
Far from mere financial engineering, the deal cements a multi-generational technology alliance: CoreWeave will deploy Nvidia’s Rubin platform, Vera CPUs, and future architectures while testing its AI-native software stack for broader ecosystem use. The goal is ambitious — more than 5 gigawatts of AI-optimized infrastructure by 2030.
Nvidia CEO Jensen Huang publicly pushed back on circular-financing critiques, while CoreWeave CEO Michael Intrator called such claims “ridiculous,” noting the investment represents just a fraction of the $25 billion-plus in total capital raised. For CoreWeave, this is validation from its largest supplier and a major customer, locking in GPU purchases and providing non-dilutive growth capital at a time when power and land costs are skyrocketing.
The partnership doesn’t just fund expansion; it future-proofs the platform against hyperscaler competition.
A New Partnership Ushers In Inference Revenue Wave
Hot on the heels of the Nvidia deal, CoreWeave announced a multi-year strategic agreement earlier this month to power Perplexity’s AI inference workloads. Using dedicated Nvidia GB200 NVL72 clusters, the arrangement delivers the low-latency, high-throughput performance essential for Perplexity’s real-time, always-on answer engine. Unlike one-off training jobs, inference represents the durable, recurring revenue phase of the AI cycle — think continuous user queries rather than occasional model builds.
CoreWeave’s bare-metal, GPU-first architecture gives it a decisive edge over generalized hyperscalers by eliminating virtualization overhead and enabling direct hardware access. This deal isn’t isolated; it signals CoreWeave’s pivot toward inference as the “new gold rush,” where specialized providers thrive.
By landing a high-profile, growth-stage AI company like Perplexity, CoreWeave diversifies its customer mix, secures sticky multi-year revenue, and demonstrates that its platform can handle production-scale demands reliably. The stock popped on the news precisely because it validates the massive capex already underway.
Real Risks That Can’t Be Ignored
CoreWeave’s growth engine requires enormous fuel. Capital expenditures are projected to surge to $30 billion to $35 billion in 2026, up from $14.9 billion last year, pushing interest expense guidance for Q1 alone into the $510 million to $590 million range. Net losses widened sharply in the fourth quarter, and free cash flow remains deeply negative as the company races to build capacity ahead of demand.
Critics rightly worry that sustained leverage, combined with potential softening in GPU pricing or utilization rates, could strain the balance sheet if AI adoption slows or power costs spike. Recent earnings triggered a sharp sell-off, underscoring investor unease about funding sustainability and execution risk in an already capital-intensive industry.
The Overshadowed Backlog
Lost in the capex panic is CoreWeave’s massive contracted revenue backlog of $66.8 billion, secured across roughly five-year average contract lengths. This pre-sold capacity — anchored by landmark commitments from OpenAI, Meta Platforms (NASDAQ:META), Microsoft (NASDAQ:MSFT), and others — essentially de-risks the spending spree. Every dollar invested today is already matched to future cash flows from long-term customers who need the GPUs now and for years to come.
Management projects an annualized revenue run rate of $17 billion to $19 billion by the end of 2026, more than tripling the current base. In short, the market’s laser focus on debt metrics has obscured the clearest proof of demand: billions in signed contracts that turn today’s infrastructure build-out into tomorrow’s predictable, high-margin revenue stream.
Key Takeaway
Investors are absolutely right to keep a close eye on CoreWeave’s debt levels and explosive capital spending; unchecked leverage in a fast-evolving AI landscape could create painful downside. Yet fixating solely on those metrics risks missing the forest for the trees.
The same growth drivers — Nvidia’s deepening partnership, inference wins like Perplexity, and a fortress-like backlog — are precisely what can service and ultimately sustain those investments as revenue scales. After the 56% slide, CoreWeave still remains premium-priced on traditional metrics, but the tailwinds from continued customer wins and AI infrastructure demand are powerful enough to warrant a serious look.
For believers in the multi-trillion-dollar AI build-out, the current discount may prove too attractive to ignore.
Top Gaining Stocks
Top Losing Stocks