Key Takeaways:
CoreWeave's multi-billion-dollar agreement with Hudson River Trading puts Nvidia's newest Vera Rubin systems to work in quantitative trading, a market the AI cloud provider is rapidly building out.
CoreWeave has signed a multi-year, multi-billion-dollar agreement with quantitative trading firm Hudson River Trading to run AI workloads on Nvidia's Vera Rubin NVL72 and HGX B200 systems, extending the cloud provider's reach into financial services. The deal, announced Aug. 20, builds on a partnership that began in March when HRT became a CoreWeave customer.
"As we scale our AI and machine learning research, the AI platform we build on matters as much as the models we build," Kevin Lee, head of research and development at HRT, said. "Giving our researchers access to the latest AI technology is exactly what lets a research team like ours push further, faster."
Under the agreement, HRT will connect its on-premises data centers to CoreWeave's cloud via high-throughput Direct Connect networking built on Nvidia's Spectrum-X Ethernet, designed to minimize latency for model training against expanding datasets. CoreWeave will also provide direct access to technical experts and a collection of AI tools and frameworks. The deal marks one of the first wide-scale deployments of Vera Rubin, Nvidia's next-generation platform, to a financial customer.
For CoreWeave, the agreement deepens a push into Wall Street that began in April when rival quant firm Jane Street committed about $6 billion in AI cloud services and invested $1 billion in CoreWeave shares at $109 apiece. Financial institutions are emerging as a critical customer segment as the company works to reduce its historical dependence on Microsoft.
Vera Rubin's First Financial Workload
Nvidia's Vera Rubin NVL72, the successor to the Blackwell architecture, pairs the Vera CPU with the Rubin GPU in a rack-scale design. CoreWeave said the clusters provide the density and throughput needed for large-scale model training, letting HRT's researchers iterate on complex models at speeds aligned with fast-moving markets. The company cited record MLPerf benchmark results and a top Platinum ranking in SemiAnalysis ClusterMAX 1.0 and 2.0 as evidence of its performance, alongside a first-place finish for inference speed and price-performance on Moonshot AI's Kimi K2.6 in testing by Artificial Analysis.
Wall Street's AI Arms Race
HRT, one of the largest nonbank trading firms in the US, has long applied machine learning and high-performance computing to quantitative research. As models grow more complex, the firm needs greater computing capacity to train against expanding datasets and iterate more rapidly. CoreWeave Chief Revenue Officer Jon Jones said financial firms are adopting the same advanced infrastructure used by frontier AI developers as trading companies expand their use of artificial intelligence. The company highlighted its growing HRT relationship in its second-quarter earnings report last week, and CoreWeave has raised billions of dollars in debt to finance additional capacity tied to demand from customers including HRT and Jane Street.
CoreWeave, which listed on Nasdaq in March 2025, trades as a specialized AI infrastructure play. The HRT deal, alongside Jane Street's $6 billion commitment, supports the case that financial services will become a durable revenue stream beyond AI labs and hyperscalers. Nvidia, meanwhile, gains another anchor customer for Vera Rubin as it competes with AMD's Instinct line and in-house chips from hyperscalers.
This article is for informational purposes only and does not constitute investment advice.
