Anthropic Partners with Google and Broadcom for Major AI Compute Capacity Deal

This agreement signals a critical shift in the competition for energy resources between AI providers and bitcoin miners.

• On April 6, 2026, Anthropic announced a multi-gigawatt TPU capacity agreement with Google and Broadcom.
• The deal, set to activate in 2027, aims to support the training of Claude AI models amid rising demand.
• This development intensifies competition for cheap electricity and resources between AI infrastructure providers and bitcoin miners.

• AI compute demand is skyrocketing, with Anthropic's annual run-rate revenue exceeding $30 billion.
• Bitcoin mining consumes 13-25 gigawatts globally, creating a race for sites with cheap power, particularly in the U.S.
• Miners are pivoting to AI hosting, seeking stable revenue as bitcoin prices fluctuate and economic pressures mount.

On April 6, 2026, Anthropic unveiled its largest compute commitment to date, securing multiple gigawatts of next-generation Tensor Processing Unit (TPU) capacity from Google, supplied by Broadcom. This deal, which will begin deployment in 2027, is designed to enhance the capabilities of Anthropic's Claude AI models, responding to a rapidly growing market that demands advanced AI solutions. With Anthropic's revenue projected to exceed $30 billion annually, the urgency for robust computing resources is palpable.

This agreement builds on Anthropic's previous $50 billion commitment to U.S. computing infrastructure made in November 2025. As AI technologies evolve, the need for powerful compute resources has surged, leading to a competitive landscape where AI infrastructure providers vie for limited energy resources. The deal is not just about securing compute capacity; it represents a strategic maneuver in a broader economic context where energy is becoming increasingly scarce and valuable.

The backdrop to this deal is the explosive growth of AI and the corresponding demand for electricity. Bitcoin mining, which consumes an estimated 13-25 gigawatts of power globally, has historically secured sites with cheap energy, especially in regions like Texas. However, as AI needs escalate, miners such as Core Scientific, Iris Energy, and Hut 8 are adapting by repurposing their facilities for AI hosting. This shift allows them to secure predictable contracts, offering a more stable revenue stream compared to the volatile nature of bitcoin prices, which have recently hovered around $69,000 with record hashrates exceeding 1 zetahash per second.

This competition for energy resources is not just a technical challenge; it has profound implications for the future of both industries. As AI providers like Anthropic ramp up their operations, the demand for electricity will likely outstrip supply, leading to potential conflicts over resource allocation. The strategic pivot of bitcoin miners towards AI hosting highlights the interconnectedness of these sectors and the necessity for both to adapt to the changing landscape of energy availability.

• AI companies: Increased energy costs and competition for resources could impact operational budgets.
• Bitcoin miners: They may need to further adapt their business models to remain profitable amid rising energy demands.
• Energy providers: They will face pressure to meet the growing demand from both sectors, potentially leading to price increases.
• Regulatory bodies: They may need to intervene to manage resource allocation and energy distribution effectively.

• Energy prices: Monitor fluctuations in electricity costs as competition between AI and bitcoin mining intensifies.
• Regulatory changes: Watch for potential regulations aimed at managing energy distribution between competing sectors.
• Market adaptations: Observe how companies in both AI and bitcoin mining adjust their strategies in response to resource scarcity.