Laika AI
Last Updated
April 7, 2026

Anthropic, one of the most closely watched artificial intelligence companies in the world, has secured multiple gigawatts of next-generation Tensor Processing Unit (TPU) capacity from tech giants Google and Broadcom, with the supply arrangement set to begin in 2027. The move reflects the company's aggressive posture in scaling compute infrastructure to meet what it describes as extraordinary and accelerating demand for its AI systems.
The deal positions Anthropic ahead of what many in the industry expect to be an intensifying compute war among frontier AI developers.
The infrastructure commitment comes at a pivotal moment for Anthropic. The company currently operates with a run-rate revenue surpassing $30 billion, a figure that places it firmly among the most commercially successful AI firms globally. Adding further weight to that trajectory, Anthropic now counts over 1,000 enterprise customers, each spending $1 million or more annually on its products and services.
These figures signal that Anthropic has successfully crossed from a research-stage AI lab into a large-scale commercial enterprise, with a customer base that includes some of the world's largest corporations across finance, healthcare, technology, and legal sectors.
TPUs, originally developed by Google, are application-specific chips purpose-built for accelerating machine learning workloads. Securing multiple gigawatts of TPU capacity is not a routine procurement decision. It represents a long-term strategic commitment tied directly to projected growth in model training, inference, and enterprise deployment at scale.
No comments yet. Be the first!
By partnering with both Google and Broadcom, Anthropic is reportedly diversifying its hardware supply chain while ensuring redundancy across next-generation chip architectures. This dual-vendor strategy reduces the risk of supply bottlenecks that have affected other AI developers in recent years.
This announcement carries significant weight beyondAnthropicitself. The scale of compute being secured sends a clear signal to the broader AI ecosystem that the race for frontier model development is entering a new, more capital-intensive phase.
Rivals, including OpenAI, Google DeepMind, and Meta AI, are similarly investing heavily in custom silicon and data center infrastructure. Anthropic's move to lock in supply through 2027 and beyond suggests the company anticipates sustained and steep growth in demand, rather than a plateau.
For enterprise buyers, this is reassuring news. Guaranteed compute capacity means Anthropic can commit to reliability, latency, and availability guarantees that large-scale enterprise contracts typically require.
Observers will be closely monitoring whether Anthropic parleys this infrastructureexpansion into new model releases, expanded API capacity, or deeper integrations with enterprise platforms. Given the company's revenue trajectory and the scale of the Google and Broadcom commitments, a major product announcement in the second half of 2026 or early 2027 would not be unexpected.
The compute deal underscores one clear takeaway: Anthropic is not slowing down.