i
News
News · 2026-10-26

Jane Street puts $7 billion behind AI compute and CoreWeave

@neuronium_ai @neuronium_ai

High-frequency trading firms are building AI infrastructure at a scale that would have looked out of place in finance just a few years ago. XTX Markets is building a data center in Finland, while Jane Street runs one in Texas with more than 4,000 Nvidia GPUs. The shift matters because these firms are not just buying access to computing: they are committing billions to providers and backing a chip startup, even as the cost and useful life of the hardware remain uncertain.

Cover: Jane Street puts $7 billion behind AI compute and CoreWeave

From cloud contracts to chip bets

One industry executive estimates that computing costs are now 10–20 times higher than before ChatGPT launched.

The companies are taking different routes to secure capacity:

Hudson River’s head of AI, Iain Dunning, said the firm spends $1 billion a year on AI and has signed a multiyear agreement with CoreWeave.
Jane Street invested $1 billion directly in CoreWeave and committed to spending $6 billion on its computing services.
Jane Street also led an investment round in chip startup Etched and became the first buyer of its AI chip, designed to deliver higher performance at lower voltage.
CoreWeaveEtched

The cost of waiting

Bloomberg reports that industry participants see rising costs and rapidly obsolete chips as serious risks. Yet executives, including Jan Bomars, CEO of Amsterdam-based high-frequency trading firm Optiver, argue that doing nothing is more dangerous.

That is the central tension: capacity is expensive, and today’s hardware may not stay useful for long. I think the striking part is not simply that trading firms want more compute; it is that Jane Street is spreading its bet across both a cloud provider and a chip maker. The announcement says little about what returns would justify those commitments, or how quickly the firms expect their AI workloads to repay them. Without that, the spending signals urgency more clearly than it demonstrates an advantage.

Daily AI news

Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.

Only what matters — every day

Follow on X