Report Details Wintermute's $1 Billion AI Investment
Wintermute reportedly plans to invest up to $1 billion over the next five years in AI data center infrastructure and high-frequency trading. The expansion beyond crypto markets targets equities, commodities and foreign exchange, with non-crypto activity expected to exceed half of the business by the end of 2027, according to reporting from Cointelegraph and Crypto Briefing based on a Bloomberg interview.
Crypto market maker Wintermute is targeting up to $1 billion over five years for high-frequency trading and AI data center infrastructure, according to Cointelegraph and Crypto Briefing, both citing reporting from Bloomberg. The reported investment accompanies a push beyond crypto trading into equities, commodities and foreign exchange.
Crypto Briefing reported that CEO Evgeny Gaevoy discussed the expansion in an interview with Bloomberg and identified established market makers Jane Street and Citadel Securities as competitors in those markets. Cointelegraph reported that Wintermute's non-crypto market activity currently accounts for 10% of its business, with a reported target of more than 50% by the end of 2027.
Compute and quantitative trading infrastructure
According to Crypto Briefing, the investment is intended for compute, storage, networking, and infrastructure used to train and continuously retrain quantitative models on large market-data volumes. The outlet also reported that Wintermute expects to finance the expansion through retained earnings, and that the firm said it remained profitable in 2025 and was on track to remain profitable in 2026.
The technical requirements for high-frequency trading differ from conventional enterprise AI deployments. Comparable trading operations typically combine low-latency data ingestion and network design with model research, backtesting, production monitoring, and rapid retraining. AI data center spending alone does not establish trading performance: market-data quality, execution infrastructure, latency controls, and risk systems are also core components of quantitative operations.
Expansion beyond crypto
Crypto Briefing reported that Wintermute's crypto trading volume had declined to about $10 billion a day from $15 billion a year earlier amid a bear market. The outlet also reported that Wintermute had entered exchange-traded funds, real-world-asset perpetual futures and prediction markets, and that its US affiliate recently registered as a broker-dealer.
Cointelegraph reported that Wintermute intends to double the staff at its 17-person New York office next year and increase global headcount by about 40%. Those hiring intentions, like the investment amount and business-mix target, are reported plans rather than completed actions.
For ML engineers and data-platform teams, the announcement illustrates a broader pattern in which quantitative finance firms treat model development and data-center capacity as linked investments. In comparable environments, continuous retraining requires governance around historical data revisions, feature reproducibility, model-version traceability, and strict separation between research systems and live execution paths.
Cointelegraph said it contacted Wintermute for additional details about the expansion. The retrieved reporting does not provide technical specifications for the proposed infrastructure, such as accelerator type, data-center location, model architectures, latency targets, or a timeline for deploying the reported capital.
Key Points
- 1Wintermute is reportedly targeting up to $1 billion for AI infrastructure and high-frequency trading over five years, expanding beyond crypto markets.
- 2Reporting puts non-crypto activity at 10% of Wintermute's business, with a stated target above 50% by the end of 2027.
- 3Comparable quantitative trading expansions require more than compute, combining low-latency data, reproducible research, continuous retraining, and execution-risk controls.
Scoring Rationale
The reported $1 billion commitment is a notable infrastructure and business expansion by a major crypto market maker into quantitative traditional-finance trading. It is relevant to ML practitioners because the reporting specifically identifies compute, storage, networking, and continuous model retraining, though no technical architecture or product release has been disclosed.
Sources
Public references used for this report.
Practice with real FinTech & Trading data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all FinTech & Trading problems
