High-frequency trading firms are expanding from software and market data into direct ownership and financing of AI infrastructure. Jane Street, Hudson River Trading and XTX Markets are spending on cloud contracts, data centers, GPUs and chip startups as machine learning becomes more central to finding trading signals.
One industry executive estimates compute spending is now 10 to 20 times higher than before ChatGPT’s launch. XTX Markets is building a data center in Finland, while Jane Street operates a Texas facility with more than 4,000 Nvidia GPUs. Hudson River Trading spends about $1 billion a year on AI and has signed a multiyear CoreWeave agreement, according to its AI chief.
Jane Street has combined supply contracts with investment. It put $1 billion into CoreWeave and committed $6 billion to the provider’s computing services. It also led a funding round for Etched and became the first customer for the chip startup, which claims its lower-voltage design can provide higher performance. Those performance claims still need to be proven in production.
The strategy reflects an arms race in which lower latency or better predictions can have immediate financial value. It also exposes trading firms to fast hardware depreciation, high power costs and uncertain returns from specialized chips. Executives cited by Bloomberg view underinvestment as the larger competitive risk, but owning compute turns infrastructure timing into part of the trading business itself.