Anthropic, the developer behind the AI chatbot Claude, appears to be shifting away from sole reliance on Microsoft Azure. A multi-year contract with cloud provider Coreweave is set to supply the ambitious startup with the substantial computing power required for training and operating the next generation of Claude. While the specifics of the agreement remain undisclosed, its very existence points to significant shifts occurring within the AI infrastructure market.
For Coreweave, a company previously heavily dependent on Microsoft, with approximately 67% of its revenue stemming from that partnership, the addition of Anthropic signifies more than just another large client. This deal offers Coreweave a crucial opportunity to diversify its business and reduce its dependence on a single major partner. Considering Coreweave's past agreements, including a reported $11.9 billion deal with OpenAI and a $21 billion deal with Meta, it is becoming evident that this specialized cloud provider is evolving into a standalone player capable of attracting highly demanding clients, irrespective of their existing relationships with technology behemoths.
This move by Anthropic is likely an attempt to gain some distance from Microsoft Azure and AWS. In the current market, where NVIDIA GPUs have become exceedingly valuable and their costs are escalating faster than AI-driven expectations, securing control over computing infrastructure is paramount. Partnering with Coreweave offers Anthropic a pathway to manage its expenditure and circumvent the pricing strategies and availability issues often associated with the major cloud providers. It is noteworthy that the solution to these challenges is emerging not from a hyperscale provider but from a niche player.
What does this development portend for the broader industry? In the ongoing AI arms race, the focus is often placed on the model developers, with less attention given to the entities that power their operations. Specialized GPU infrastructure providers like Coreweave are transitioning from the role of "technical contractors" to becoming critical players in the ecosystem. They are creating viable alternatives for AI startups, enabling them to avoid the trap of dependency on cloud monopolists. For business leaders, this serves as a clear signal: the cost of AI computation may become increasingly unpredictable, and strategic partnerships with hardware suppliers could prove more valuable than simply utilizing traditional cloud services.