Anthropic is in advanced talks with Nscale, a cloud computing startup backed by Nvidia, for a deal valued at approximately $45 billion, Reuters reported on Thursday, citing a source familiar with the negotiations. If completed, the agreement would be one of the largest AI infrastructure contracts ever struck, underscoring the extraordinary scale of compute capacity that frontier AI labs now require to train and deploy their models.
A Pattern of Large-Scale Compute Deals
The reported Nscale talks fit a clear pattern for Anthropic, which has been aggressively assembling a diversified portfolio of compute partnerships over the past year. The company previously struck a $10 billion compute arrangement with another Nvidia-backed startup, a deal that signaled its appetite for infrastructure at a scale that few companies outside the hyperscalers have historically pursued. That earlier agreement set the stage for increasingly ambitious negotiations across multiple hardware and cloud providers.
Key Facts
- Reported deal value: $45 billion with Nscale
- Nscale is backed by Nvidia, giving it access to high-demand GPU hardware
- Negotiations are ongoing and no final agreement has been announced
- The deal would be among the largest AI cloud contracts on record
- Anthropic has pursued multiple large compute partnerships in parallel
Nscale, which operates GPU clusters optimized for AI workloads, has positioned itself as an alternative to the major hyperscalers for companies that need dedicated, high-throughput infrastructure. Nvidia's backing gives the startup preferential access to hardware that remains in short supply across the industry, a factor that makes it an attractive partner for a company like Anthropic, which is racing to scale its Claude model family across enterprise and consumer markets alike.
The scale of the reported deal suggests Anthropic is planning for a compute footprint that would rival the internal infrastructure budgets of some of the world's largest technology companies.Reuters
Expanding Beyond Any Single Cloud Provider
Anthropic's infrastructure strategy has increasingly moved toward distributing its compute across multiple vendors rather than concentrating capacity with a single provider. Earlier this year, the company signed a $1.8 billion deal with Akamai to expand inference capacity for Claude, and separate reports have detailed discussions around a potential $10 billion arrangement with Meta for computing resources. The Nscale talks, if they result in a signed contract, would add another major pillar to that architecture.
The financial scale of these agreements also reflects how quickly Anthropic's business has grown. The company's revenue run rate has climbed sharply as enterprise adoption of Claude accelerates, and sustaining that trajectory requires a continuous build-out of the underlying hardware that powers model training and inference. A $45 billion commitment, even spread across several years, represents a bet that demand for Anthropic's services will continue to grow at a substantial rate. For the broader picture of Anthropic's compute and revenue growth, the Nscale talks are the latest indicator that the company is planning for a future where its infrastructure requirements are measured in the tens of billions.
Neither Anthropic nor Nscale has publicly confirmed the reported negotiations. The terms of the deal, including the timeline and structure of any payments, have not been disclosed. Talks of this size often evolve significantly before any final agreement is reached, and it is possible the reported figure or scope could change. Regardless of the outcome, the fact that discussions at this scale are taking place speaks to how much the competitive dynamics of the AI industry have shifted in a short period of time.
