The artificial intelligence race is no longer being won by algorithms alone. Instead, it is increasingly being decided by who can secure the most powerful chips, the largest data centers, and the biggest pools of electricity. Anthropic’s reported $10 billion computing agreement with AI infrastructure startup Volta Infra Holdings is the latest example of this dramatic shift.
The deal, one of the largest AI infrastructure agreements announced this year, underlines a simple truth: in today’s AI economy, access to computing power has become as valuable as access to talent or capital.
Credits: Bloomberg
Anthropic Locks in Massive Computing Capacity
According to reports, Anthropic has signed a six-year agreement to use computing capacity managed by Volta Infra Holdings, a cloud infrastructure startup backed by Nvidia. Although neither company officially confirmed the partnership, Volta announced that it had secured a $10 billion contract with an unnamed AI laboratory, with infrastructure to be delivered alongside Bitcoin mining company Bitdeer at a data center in Norway.
The Norwegian facility will provide 133 megawatts of computing capacity and will be equipped with Nvidia’s next-generation Vera Rubin AI chips, which are expected to power some of the world’s most advanced AI models.
For Anthropic, this isn’t simply another cloud contract—it is an investment in ensuring that Claude, its flagship AI assistant, has enough computational resources to meet rapidly growing demand from enterprises and consumers.
AI Companies Are Racing to Secure Compute
Training and running frontier AI models has become extraordinarily expensive. Every new generation of AI requires larger datasets, more sophisticated models, and significantly more graphics processing units (GPUs).
As demand explodes, AI companies are rushing to reserve computing capacity years in advance.
Anthropic has already signed infrastructure agreements with companies including SpaceX, AMD, and Akamai Technologies. Reports also suggest that the company is exploring leasing additional capacity from Meta’s data centers.
Its biggest rival, OpenAI, is pursuing a similar strategy. CEO Sam Altman has repeatedly argued that AI will require investments measured in trillions of dollars, not billions. OpenAI is reportedly planning a data center in Georgia that could cost over $30 billion, while discussions continue around an even larger AI campus backed by SoftBank.
The message is clear: whoever controls the infrastructure may control the future of AI.
Meet Volta: The Startup Building AI’s Backbone
Interestingly, Volta Infra Holdings was founded only this year by former Brookfield Asset Management executives.
Rather than developing AI models, Volta focuses on financing and leasing AI infrastructure. It helps AI companies gain access to expensive Nvidia chips without purchasing every piece of hardware outright.
The company also announced a $300 million funding round, giving it a valuation of $2.4 billion only months after its launch.
That investors are willing to value such a young company so highly reflects growing confidence that infrastructure providers could become some of the biggest winners of the AI boom.
Credits: Bloomberg
Bitcoin Miners Find a New Business Model
The agreement also illustrates another fascinating industry trend.
Volta’s infrastructure partner, Bitdeer Technologies, originally built its data centers to mine Bitcoin. However, with cryptocurrency mining becoming less profitable amid weaker Bitcoin prices, companies like Bitdeer are increasingly converting their facilities into AI computing hubs.
Instead of solving cryptographic puzzles, these massive facilities will soon be training large language models and supporting AI applications.
The transition demonstrates how existing digital infrastructure is being repurposed to serve one of the fastest-growing technology markets in history.
The Growing Risks Behind the AI Boom
While the scale of these investments is impressive, analysts have also raised concerns.
Many AI companies are funding infrastructure through increasingly interconnected partnerships involving cloud providers, chip manufacturers, investors, and AI developers. Critics argue that this creates a complex web of financial dependencies.
If demand for generative AI fails to justify these enormous investments, the impact could ripple across the broader technology ecosystem, affecting infrastructure providers, semiconductor companies, and cloud operators alike.
For now, however, confidence remains remarkably strong.
Credits: Reuters
The Future Will Be Built on Compute
Anthropic’s reported $10 billion agreement is about much more than renting servers. It represents the new reality of artificial intelligence, where success depends on securing long-term access to the world’s most advanced computing infrastructure.
As AI models continue to grow larger and more capable, computing power has become the industry’s most valuable resource. Companies are no longer competing solely to build better AI—they are competing to build the digital factories that make AI possible.
If data was the fuel that powered the first wave of artificial intelligence, compute is rapidly becoming the foundation of the next.