Anthropic's $10 Billion Leap: Partnering with AI Cloud Startup Volta
TL;DR
- Anthropic has reportedly struck a $10 billion, six-year computing deal with Volta Infra Holdings, a new AI cloud startup backed by Nvidia, to secure more capacity for training and serving Claude models.
- The compute will be delivered through a Norway data center operated with Bitdeer, using Nvidia Vera Rubin AI chips and about 133 MW of capacity.
- The deal underscores a broader industry shift: AI labs are racing to lock in scarce compute through long-term cloud and infrastructure partnerships.
Anthropic’s latest infrastructure move is a major signal that the battle for AI compute is intensifying. According to multiple reports, the Claude maker has agreed to spend about $10 billion over six years for computing capacity from Volta Infra Holdings, a months-old, Nvidia-backed cloud startup.
A big bet on compute
The reported agreement is not just a routine cloud contract; it reflects the scale of demand behind frontier AI systems. Anthropic has been expanding capacity to support growing usage of Claude, especially for coding and other high-intensity workloads, and the Volta deal appears designed to keep that growth from being constrained by infrastructure shortages.
Volta itself is a fresh entrant in the market. Reuters reported that the company was founded in January by former Brookfield Asset Management executives and recently raised $300 million in seed and Series A financing, valuing it at $2.4 billion. Its backers include Azora, Andreessen Horowitz, Altimeter, and Nvidia.
Why Norway matters
The compute behind the deal is expected to come from a data center in Norway, developed with Bitdeer Technologies Group, a bitcoin-mining firm that operates data centers. The site is described as having roughly 133 megawatts of capacity and being powered by Nvidia’s Vera Rubin systems.
That geography is significant because Norway offers access to large-scale energy infrastructure and cooler operating conditions, both attractive for power-hungry AI workloads. The deal also illustrates a growing pattern in which crypto-related infrastructure is being repurposed or extended into AI hosting capacity.
Volta’s fast rise
Volta is moving unusually quickly for a startup. Beyond the Anthropic-linked contract, the company also announced a separate $5 billion AI infrastructure program with asset manager Azora to finance future data center projects.
Reuters noted that the Norway site is only the first step in a broader development pipeline that Volta says could exceed one gigawatt of near-term power capacity across North America and Europe. That would place the company in the league of major infrastructure players far sooner than typical cloud startups.
Anthropic’s broader cloud strategy
The Volta arrangement fits a larger pattern: Anthropic has been actively diversifying and deepening its compute relationships. Reuters reported that the company has signed a series of computing and investment agreements with firms including SpaceX, AMD, Microsoft, Nvidia, Alphabet, Amazon, and Micron.
That strategy suggests Anthropic is trying to avoid overreliance on any single supplier while ensuring enough capacity to train and run increasingly demanding models. In the current AI market, access to compute is becoming as strategically important as model quality itself.
What this means for the AI industry
The deal reinforces a central trend in 2026: AI companies are no longer just buying cloud services, they are effectively locking up industrial-scale infrastructure through long-duration partnerships.
It also highlights how Nvidia’s ecosystem has become deeply embedded in AI expansion. Volta is backed by Nvidia and the Norway deployment is expected to use Nvidia’s next-generation chips, showing how chipmakers, startups, and model developers are increasingly linked in the same capital and infrastructure loop.
The unanswered questions
Several details remain unconfirmed publicly. Anthropic and Bitdeer declined to comment, and Volta did not publicly name its customer when it announced the deal. Bloomberg’s identification of Anthropic came through people familiar with the matter, not a direct company confirmation.
Even so, the reported terms are large enough to matter on their own. If finalized as described, the agreement would be one of the most consequential compute commitments in the current AI boom, with implications for how frontier model labs plan capacity, financing, and global infrastructure footprints.
Get All The Latest Updates Delivered Straight To Your Inbox For Free!