Crusoe Signs a $13 Billion AI Cloud Deal With Jane Street — and Nears a $30B Valuation

The AI infrastructure land grab has a new headline number: $13 billion. According to Bloomberg News, San Francisco-based Crusoe has signed a five-year cloud contract worth roughly $13 billion with Jane Street Group, the quantitative trading firm that has quietly become one of the largest private buyers of AI compute on the planet. Reuters, citing the same report, notes that the deal will supply Jane Street with clusters of GPUs and other infrastructure needed for AI training and inference through Crusoe’s cloud platform.

Neither Crusoe nor Jane Street immediately responded to requests for comment, and key details — which GPU systems Jane Street will receive, how much capacity the contract covers, and when it comes online — remain undisclosed. But the headline figure alone makes Jane Street the highest-profile cloud customer Crusoe has landed to date, and it lands at a moment when the “neocloud” sector is consolidating around a handful of scaled players.

What we know about the deal

The contract, reported Thursday, September 3, spans five years and is valued at approximately $13 billion. Under its terms, Crusoe will provide Jane Street with GPU clusters and supporting infrastructure for both training and inference workloads via Crusoe Cloud. For a trading firm, that dual use matters: Jane Street’s machine learning workloads span everything from research model training to latency-sensitive production inference across global markets.

The scale is easier to grasp in context. Jane Street already operates thousands of GPUs in its own data centers — its newest Texas facility alone houses 4,032 GPUs across 56 racks. In April, the firm committed approximately $6 billion to CoreWeave’s AI cloud platform and made a separate $1 billion equity investment in that company, gaining access to compute across multiple facilities, including infrastructure based on NVIDIA’s Vera Rubin technology. Add the reported Crusoe agreement, and Jane Street’s combined external cloud commitments now approach $19 billion — before counting its owned capacity.

The fundraise that followed

The Jane Street contract did more than add a marquee logo to Crusoe’s customer list — it supercharged the company’s fundraising. Bloomberg reports that the deal helped attract fresh interest in Crusoe’s latest round, and the numbers now circulating are striking: Crusoe has closed more than $3 billion at a post-money valuation of roughly $30 billion, with Atreides Management and Valor Equity Partners co-leading and Mubadala Capital participating.

If those figures hold, the round nearly triples Crusoe’s valuation from its October 2025 Series E, when it raised $1.375 billion at a valuation above $10 billion, co-led by Mubadala Capital and Valor. A 3x markup in under a year is aggressive even by AI-infrastructure standards — but it tracks a sector where revenue visibility, not current earnings, drives valuations. A five-year, $13 billion contract with an investment-grade counterparty is precisely the kind of predictable cash flow that lenders and equity investors price favorably.

Why Crusoe, and why now

Crusoe has spent the past two years positioning itself as more than a GPU reseller. The company describes itself as an “AI factory” operator with a vertically integrated stack: it develops energy infrastructure, builds large-scale data centers, and sells GPU compute through Crusoe Cloud. As of June, Crusoe said it had contracted 4.9 gigawatts of AI infrastructure capacity across its data center projects and cloud business.

Power, not chips, is often the binding constraint on data center expansion, and Crusoe’s approach — combining grid interconnections with on-site and behind-the-meter generation — is designed to bring facilities online faster in markets where local grid capacity is constrained. That capability increasingly differentiates neoclouds as hyperscalers and enterprises compete for the same megawatts.

The competitive backdrop also matters. Jane Street’s earlier $6 billion CoreWeave commitment showed that financial firms would sign nine-figure-per-year contracts with specialized cloud providers rather than wait in hyperscaler queues. The reported Crusoe deal suggests those buyers are now deliberately multi-sourcing — spreading commitments across CoreWeave, Crusoe, and their own facilities to hedge supply risk, pricing, and technology transitions.

The bigger picture: finance is eating the compute market

The most underappreciated story in AI infrastructure may be how quickly quantitative finance has become a first-tier compute customer. Jane Street is now simultaneously an equity investor in CoreWeave, the lead investor in transformer-chip startup Etched’s recent rounds, the lead in Fluidstack’s $1.5 billion round at an $18 billion valuation (reported just this week), and — if Bloomberg’s report is accurate — a $13 billion cloud customer of Crusoe.

That pattern reflects a simple reality: trading firms monetize prediction advantages at enormous scale, and frontier AI models are prediction machines. A firm that can shave milliseconds or improve forecasts by fractions of a percent can plausibly earn returns that justify billions in annual compute spend. Wall Street’s internal AI buildouts are no longer a side project — they are a demand curve that rivals, and in some cases exceeds, that of AI labs.

For the neocloud sector, the message is equally clear. Crusoe’s reported week — a $13 billion contract followed by a $30 billion valuation — mirrors the trajectory that carried CoreWeave to a landmark IPO and Fluidstack to a near-tripled valuation in four months. Specialized AI cloud providers with secured power, credible operations, and anchor customers are commanding valuations that legacy data center operators never approached.

What to watch

Several open questions will determine whether this deal marks a peak or a plateau. First, confirmation: at publication time, neither company had publicly verified the contract’s existence or terms. Second, the GPU generation mix — in a market transitioning to NVIDIA’s Rubin-class systems, whether Jane Street’s clusters ship on current or next-generation silicon materially affects the deal’s economics. Third, the fundraising’s final form: Bloomberg’s July report noted a final valuation had not been set, and terms could shift before any announcement.

What is already clear is the direction. AI compute has become a strategic input for the world’s most sophisticated financial firms, and suppliers who can guarantee power and GPUs at scale are being repriced accordingly. Crusoe’s reported double announcement — a $13 billion customer and a $30 billion valuation — is the latest, and one of the largest, data points in that repricing.