Anthropic has signed a monumental $45 billion agreement with data center operator Nscale, marking its latest and largest move to secure the vast computing resources needed to train and run advanced AI models. The deal, reported by TechCrunch, represents a multi-year commitment and underscores the intense, capital-heavy competition for specialized AI infrastructure.
This agreement continues what the report describes as a "compute-gobbling streak" for the AI safety startup. It follows closely on the heels of Anthropic's other massive infrastructure deal: a $15 billion annual commitment to SpaceX for compute capacity, as detailed in a separate analysis. That arrangement, valued at $1.25 billion per month, highlighted the extreme lengths to which AI firms are going to bypass traditional cloud providers and secure dedicated hardware.
The choice of Nscale as a partner is significant, given the data center company's own high-profile backing. Earlier this year, Nvidia CEO Jensen Huang personally invested £500 million in Nscale, a move framed as a symbolic bet on the future of AI infrastructure. Huang reportedly predicted that Nscale could reach $50 billion in revenue, indicating strong confidence in the company's ability to scale and meet the demands of clients like Anthropic.
Anthropic's aggressive compute acquisition strategy is driven by the astronomical processing requirements of cutting-edge large language models and the vision of a vast commercial opportunity. According to a separate report, the company is preparing to tell investors that it sees a total addressable market (TAM) exceeding $30 trillion. This figure would surpass even SpaceX's recently cited $28.5 trillion TAM and signals Anthropic's belief that enterprise applications of AI alone could capture nearly $23 trillion in value.
The scale of these infrastructure deals reflects a fundamental shift in the AI industry. As models grow larger and more complex, access to specialized compute—primarily powered by Nvidia's advanced GPUs—has become the critical bottleneck and competitive moat. Companies are no longer just competing on algorithmic innovation but on their ability to secure and fund the physical hardware required for training and inference. This dynamic favors well-funded players and creates intense partnerships between AI labs, chipmakers, and infrastructure providers.
Anthropic's dual deals with SpaceX and Nscale suggest a multi-pronged approach to solving its compute needs, potentially diversifying its supply chain across different types of infrastructure providers. The Nscale deal, valued in the tens of billions, represents a traditional yet massive data center commitment, while the SpaceX partnership may involve more novel or specialized capacity. Together, they represent a staggering ongoing financial commitment, underscoring the belief that future AI leadership is inextricably linked to control over computing power.
The financial implications are profound. Commitments of this magnitude—$45 billion to Nscale and $15 billion annually to SpaceX—will shape Anthropic's capital requirements and valuation expectations for years to come. They also set a new benchmark for what constitutes necessary investment in the AI sector, potentially raising the barrier to entry for other firms and intensifying the race for capital among infrastructure providers.
This series of moves positions Anthropic not just as an AI research lab, but as a major force in the global scramble for compute resources. The company's actions highlight a central tension in the industry's growth: the pursuit of increasingly powerful and potentially transformative AI models is colliding with the physical and economic realities of the hardware required to build them. How Anthropic manages and funds these commitments will be a key determinant of its ability to compete against rivals like OpenAI, Google, and Meta in the coming years.
