Trading firms pour billions into neocloud infrastructure

Jane Street and rivals are betting big on GPU power to fuel their AI-driven trading strategies.

11 cards · 2m · tap to begin

From · · 2m

Trading firms pour billions into neocloud infrastructure

Jane Street and rivals are betting big on GPU power to fuel their AI-driven trading strategies.

In brief

Jane Street and rivals are betting big on GPU power to fuel their AI-driven trading strategies. High-frequency trading firms are driving real demand for neocloud infrastructure with multi-billion dollar commitments to power their AI trading models. Originally reported by Futuriom.

Trading firms fuel neocloud boom

Large U.S. trading firms are backing neoclouds with real money to secure GPU infrastructure for their complex AI workloads that power Wall Street profits.

Jane Street's $13 billion Crusoe deal

Jane Street signed a $13 billion, five-year contract with Crusoe, valuing the neocloud at about $30 billion.

CoreWeave gets $7 billion commitment

CoreWeave gets $7 billion commitment

Jane Street committed $1 billion to CoreWeave in April 2026, with a pledge to spend about $6 billion on its infrastructure.

Fluidstack and Firmus investments

Jane Street led a $1.5 billion round for Fluidstack, valuing it over $18 billion, and joined a $2 billion round for Australian neocloud Firmus.

Other trading firms join in

Hudson River Trading, Susquehanna International Group, and IMC Trading are among other high-frequency firms investing in neoclouds like CoreWeave and IREN.

Real demand vs speculation

Trading firm investments are tangible demand for datacenter expansion, unlike speculative bets such as Anthropic's $35 billion deal with Lambda.

Research labs with trading desks

Jane Street operates like a research lab, with teams inventing their own AI architectures and optimizations specifically designed for trading.

“Think of Jane Street as a research lab with a trading desk attached to it”

Sub-millisecond accuracy required

Jane Street and HRT use AI for arbitrage through Exchange Traded Funds, a process requiring extreme sub-millisecond accuracy.

Custom hardware for live execution

Jane Street designs its own hardware workflows using Hardcaml to make FPGA and ASIC-based hardware more functional for live trading execution.

Jane Street's own GPU datacenter

Beyond neoclouds, Jane Street built a 4,032-GPU datacenter in Texas to support its AI efforts alongside hyperscaler services.

Revenue justifies the spending

Jane Street scored $16.1 billion in trading revenue for Q1 2026, up over 40% year-on-year, justifying massive AI infrastructure investments.

Here's the gist

High-frequency trading firms are driving real demand for neocloud infrastructure with multi-billion dollar commitments to power their AI trading models.

Read the original on Futuriom

React

Sign in to react and comment.

Comments (0)

Life is short. Keep it sweet. Respect others' opinions and be kind!

    Recommended next

    More decks on jane street and related topics.