Jane Street signs $6 billion AI cloud deal with CoreWeave, invests $1 billion in equity
Short summary
Jane Street agreed to spend $6 billion on AI cloud compute from CoreWeave and took a $1 billion equity stake at $109 per share, making it one of CoreWeave’s top five shareholders. The deal secures access to NVIDIA’s next-gen Vera Rubin GPUs and aligns Jane Street’s trading firm with the rapid expansion of specialized AI infrastructure.
- Jane Street is spending $6B on CoreWeave AI compute and taking a $1B equity stake at $109/share, becoming a top-five shareholder—valuing CoreWeave 176% above its IPO price.
- Trading firms are now becoming major direct AI infrastructure players, not just users of it, with Jane Street's $13B net income last year making the spend easily affordable.
- CoreWeave has locked in massive deals beyond this one (Meta $35B, OpenAI $12B, NVIDIA $6.3B), positioning it as a specialized alternative to AWS/Azure for AI workloads.
- Long-term compute commitments this large carry real risk if AI demand slows or cheaper architectures emerge, turning bargains into burdens for both sides.
Full summary
Jane Street has agreed to spend $6 billion on AI cloud compute with CoreWeave and taken a $1 billion equity stake at $109 a share. That purchase makes the quantitative trading firm one of CoreWeave’s five biggest shareholders and values CoreWeave 176 percent above its IPO price just over a year ago. Under the deal, Jane Street will tap into NVIDIA’s upcoming Vera Rubin GPU systems across multiple data centers, bolstering its own fleet of high-end processors for training trading models on massive, noisy market data in real time.
CoreWeave’s rise has been rapid. It went public in March 2025 at $40 a share, and since then has signed deals dwarfing its IPO valuation. Meta has committed $35 billion through 2032, OpenAI about $12 billion over five years, and NVIDIA agreed to buy $6.3 billion of untapped capacity while also investing $2 billion directly. By focusing solely on AI workloads, CoreWeave can optimize hardware, connectivity, and support in ways AWS or Azure don’t, making it a go-to partner for customers that need maximum performance.
This arrangement shows how finance groups are becoming AI players themselves, buying compute at scale and even backing the providers. Jane Street made $13 billion in net income last year—enough to train frontier language models several times over—so spending $6 billion on cloud capacity isn’t a stretch. But it also exposes both sides to risk: if demand for large-scale AI compute slows or a new architecture cuts costs dramatically, those massive, long-term commitments could turn into burdens rather than bargains.