Jane Street is building its own AI infrastructure, a move that signals how far quantitative trading firms will go to gain an edge in machine-learning-driven markets.
Jane Street is building its own AI infrastructure, a move that signals how far quantitative trading firms will go to gain an edge in machine-learning-driven markets.

Jane Street, the quantitative trading giant, is planning to build and finance its own data center with as much as 200 megawatts of capacity, people with knowledge of the matter said, as the firm pushes deeper into artificial intelligence to sharpen its trading edge.
"The facility would be primarily for internal use, potentially to train proprietary AI models for forecasting asset prices," one of the people said, asking not to be identified discussing private information.
The firm has been in discussions with companies across technology, cryptocurrency and financial sectors about the project, which would add 100 MW to 200 MW of computing capacity on top of its existing investments in AI compute. Jane Street currently sources part of its computing power from a data center in Dallas and has partnerships with cloud providers including CoreWeave Inc. The talks remain at an early stage, with the exact capacity and location yet to be determined.
The approach differs from how private equity typically invests in AI data centers, as Jane Street would own and operate the facility primarily for its own trading operations rather than leasing capacity to third parties. Founded in 2000, the firm executes trades for investors and with its own capital, processing thousands of transactions within seconds while also holding positions for hours, days or even weeks. Building proprietary AI compute infrastructure could give Jane Street a latency and modeling advantage over rivals who rely on shared cloud resources.
The move shows a broader trend of financial firms bringing AI infrastructure in-house as machine learning becomes central to trading strategies. Unlike traditional hedge funds that lease cloud capacity, Jane Street's buildout mirrors the approach of hyperscale cloud providers that design custom infrastructure for their workloads.
A 200 MW Bet on Proprietary AI
A 200 MW data center would rank among the larger single-tenant facilities built by a non-tech company. For context, 200 MW can power roughly 170,000 US homes. The scale shows that Jane Street views AI model training as a core competitive function rather than an outsourced IT expense.
The firm's existing partnership with CoreWeave, a cloud provider specializing in GPU-accelerated workloads, suggests Jane Street has already been experimenting with large-scale AI compute. Building its own facility would give the firm greater control over hardware configuration, cooling efficiency and network latency — all critical factors for high-frequency trading strategies where microseconds matter.
What It Means for the Data Center Industry
Jane Street's move adds to the surging demand for AI-ready data center capacity. Hyperscalers including Microsoft Corp., Amazon.com Inc. and Alphabet Inc. have committed tens of billions of dollars to expand their data center footprints, while Meta Platforms Inc. has adopted rapid-deployment strategies such as building in tent-like structures to accelerate timelines.
For data center developers and equipment suppliers, the emergence of large financial firms as direct owners of compute infrastructure represents a new demand vector. Unlike cloud providers that spread capacity across thousands of tenants, a proprietary facility for a single trading firm requires specialized design for low-latency interconnects and high-density GPU clusters.
Jane Street's pre-leasing of more than 70 percent of the office space in Phase 1 of Henderson Land Development Co.'s Central Harbourfront project in June last year — the largest single office leasing transaction in Hong Kong's Central district in decades — shows the firm's willingness to make large-scale infrastructure commitments.
The buildout suggests that proprietary AI compute is becoming a competitive necessity for top-tier trading firms, potentially driving a new wave of demand for data center construction, networking equipment and GPU hardware. Companies supplying data center infrastructure — from power management systems to high-speed interconnects — could benefit as financial firms follow Jane Street's lead. However, the project remains in early stages, and the final capacity and location have yet to be determined.
This article is for informational purposes only and does not constitute investment advice.