CoreWeave signs multibillion-dollar AI cloud deal with Hudson River Trading
- CoreWeave secured a multibillion-dollar multiyear AI cloud deal with Hudson River Trading, described as a material expansion of an existing partnership
- HRT will access Nvidia Vera Rubin NVL72 and B200 systems; B200 chips currently rent for $6.87 per hour
- HRT posted record $11.4 billion in second-quarter trading revenue and $7.4 billion in profit
- CoreWeave raised prices approximately 25% in July and expects newer contracts to deliver contribution margins 5 to 10 percentage points higher than earlier agreements
- CoreWeave's largest customer accounted for 36% of second-quarter sales, down from 71% a year earlier, reflecting revenue diversification

*this image is generated using AI for illustrative purposes only.
CoreWeave Inc. has secured a multibillion-dollar multiyear agreement with Hudson River Trading (HRT), intensifying Wall Street's race for the newest AI chips from Nvidia Corp. CoreWeave Chief Revenue Officer Jon Jones called the deal "a material expansion" of an existing partnership.
Partnership details
Under the agreement, HRT will gain access to Nvidia's Vera Rubin and B200 systems through CoreWeave's purpose-built AI cloud platform. CoreWeave says it was the first AI cloud provider to validate Nvidia's Vera Rubin NVL72 platform, supported by Spectrum-X Ethernet networking. The infrastructure is designed to enable HRT to scale its AI-driven trading research and model development capabilities.
Kevin Lee, Hudson River's head of research and development, said giving researchers access to the latest AI technology is what "lets a research team like ours push further, faster."
| Deal parameter | Details |
|---|---|
| Deal value | Multibillion-dollar |
| Duration | Multiyear |
| Infrastructure | Nvidia Vera Rubin NVL72, B200 systems |
| Networking | Spectrum-X Ethernet |
Wall Street's AI arms race
HRT uses AI computing power to train trading models, test strategies, and search vast datasets for profitable patterns. The firm has significant financial resources to support this investment, having generated a record $11.4 billion in second-quarter trading revenue and $7.4 billion in profit. The deal follows rival Jane Street's $6 billion commitment to CoreWeave's AI cloud and a separate $1 billion investment in the company. CoreWeave has also added trading firms Flow Traders and IMC as customers.
Nvidia chips give CoreWeave pricing power
HRT will also use Nvidia B200 chips, which currently rent for $6.87 per hour according to the Ornn index. CoreWeave raised prices approximately 25% in July and expects newer contracts to deliver contribution margins 5 to 10 percentage points higher than earlier agreements.
| Metric | Details |
|---|---|
| B200 rental rate | $6.87 per hour |
| Price increase (July) | ~25% |
| Margin uplift on new contracts | 5 to 10 percentage points higher |
Powering growth and reducing customer risk
Meeting demand requires securing sufficient electricity to expand CoreWeave's data centers. The company finished the second quarter with approximately 3.7 gigawatts of contracted power capacity.
The agreement also comes as CoreWeave works to diversify its revenue base. Its largest customer accounted for 36% of second-quarter sales, compared with 71% for its biggest customer a year earlier. Its top three customers still generated approximately 72% of revenue.
| Customer concentration metric | Latest quarter | Prior year |
|---|---|---|
| Largest customer share of revenue | 36% | 71% |
| Top three customers share of revenue | ~72% | Not disclosed |
How might CoreWeave's 25% price hike and increased margins impact the competitive landscape among other AI cloud providers like AWS or Azure?
What are the potential risks for CoreWeave if its top three customers continue to account for approximately 72% of its revenue despite recent diversification efforts?
How will the high demand for Nvidia Vera Rubin and B200 chips affect supply chain constraints and pricing power for other financial institutions seeking similar AI infrastructure?

































