AivexaNewsSearch
AI news for builders and product teamsChecked every hour

Wall Street's fastest traders are building data centers and backing chip startups to stay ahead in AI

Collected Oct 26, 2026

Major high-frequency trading firms are redirecting large sums into AI computing infrastructure, according to a Bloomberg report summarized by The Decoder. Among the named players are Jane Street, Hudson River Trading, XTX Markets and Amsterdam-based Optiver. The spending spans cloud contracts, investments in chip startups and the construction of their own data centers. One executive estimates that compute spending across such firms is now ten to twenty times higher than it was before ChatGPT launched.

The concrete figures are substantial. Jane Street operates a data center in Texas with more than 4,000 Nvidia GPUs. XTX Markets is building a data center in Finland. Hudson River Trading spends $1 billion a year on AI, according to its AI chief Iain Dunning, and has signed a multiyear agreement with CoreWeave, a cloud provider that rents out GPU capacity.

Jane Street took a different path with CoreWeave: rather than only renting capacity, it invested $1 billion directly in the company and committed $6 billion to its computing services. That combination of equity stake plus long-term consumption contract is a pattern worth watching, since it gives the trading firm influence over a supplier while locking in access to scarce GPUs.

Jane Street also led a funding round for chip startup Etched and became the first customer for its AI chip. The chip is described as intended to deliver more performance by running at a lower voltage. In practice, lower-voltage operation is a familiar lever in semiconductor design: reducing voltage tends to cut power draw, which matters for dense data center deployments where electricity and cooling are real constraints. Whether it translates into better throughput for the workloads these firms run is something the market will test as the chip reaches customers.

Underlying the spending is a shift in how these firms trade. They increasingly rely on machine learning to predict market moves, which raises demand for compute. High-frequency trading has long been a compute-intensive business, but historically the emphasis was on low-latency networking and specialized hardware for executing orders quickly. Machine learning workloads add a different profile: training and inference on large models, which favors GPU clusters and can be run less latency-sensitively than order routing.

The article notes that the industry sees rising costs and fast-aging chips as genuine risks. GPUs depreciate quickly as new generations arrive, so large capital commitments can become obsolete faster than the facilities housing them. Still, Optiver CEO Jan Boomaars is quoted as believing that doing nothing is the more dangerous option, a signal that firms view compute access as a competitive necessity rather than an optional investment.

Why it matters: Trading firms are becoming significant buyers of AI infrastructure, competing with technology companies for GPUs, cloud capacity and data center space. For developers and product teams, that means tighter supply and pricing pressure in GPU cloud markets, while vendors like CoreWeave gain anchor customers with deep pockets. A likely inference is that specialized AI chip startups will find early customers among latency- and cost-sensitive financial firms, giving new hardware a demanding first proving ground.

Read at The Decoder

Based on reporting from the original publisher. Visit the source for full context and later updates.

Publisher excerpt

High-frequency traders like Jane Street and Hudson River Trading are now racing each other for AI computing power. The article Wall Street's fastest traders are building data centers and backing chip startups to stay ahead in AI appeared first on The Decoder .