# Anthropic announces expanded Google and Broadcom compute agreement [Canonical reading](https://escapemeridian.com/readings/anthropic-announces-expanded-google-and-broadcom-compute-41e47cf2) Anthropic announced the agreement on April 6, 2026. It says the deal with Google and Broadcom is for multiple gigawatts of next-generation TPU capacity, expected to come online starting in 2027, to power Claude models and serve customer demand. Anthropic CFO Krishna Rao describes it as part of a “disciplined approach to scaling infrastructure” and says the company is making its “most significant compute commitment to date.” Anthropic ties the expansion to rapid growth in customer demand: it reports run-rate revenue above $30 billion, up from approximately $9 billion at the end of 2025, and says the number of business customers spending over $1 million annually has doubled since its February Series G announcement. The announcement says most of the new compute will be in the United States, extending Anthropic’s earlier commitment to invest in American computing infrastructure. Anthropic also presents its use of AWS Trainium, Google TPUs and NVIDIA GPUs, alongside access to Claude through three major cloud platforms, as a way to match workloads to hardware and provide customers with better performance and resilience. The scale of capacity that will come online, and how much it will improve service, remain unspecified. Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims. ## Original source Anthropic · Article Original source date: Apr 6, 2026 (2026-04-06T00:00:00.000Z). Date basis: Date stated on the publisher’s page. Added to Escape Meridian: 2026-11-07T15:18:23.475Z. First observed: 2026-10-19T04:45:24.603Z. Retrieved: 2026-10-28T13:53:49.638Z. - [Anthropic expands Google and Broadcom compute deal](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Limits Text as retrieved on October 28, 2026; originally published April 6, 2026. Anthropic does not specify the exact amount of capacity or how much will come online at each stage. ## Open question How much of the announced capacity will come online starting in 2027, and what effect will it have on Claude service? ## Source-stated Anthropic says its agreement with Google and Broadcom covers multiple gigawatts of next-generation TPU capacity, expected to come online starting in 2027. > We have signed a new agreement with Google and Broadcom for multiple gigawatts of next-generation TPU capacity that we expect to come online starting in 2027. - [Supporting source](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Source-stated Anthropic CFO Krishna Rao calls the expansion part of a disciplined approach to scaling infrastructure and its most significant compute commitment to date. > “We are making our most significant compute commitment to date to keep pace with our unprecedented growth.” - [Supporting source](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Source-stated Anthropic reports that its run-rate revenue surpassed $30 billion, compared with approximately $9 billion at the end of 2025. > Our run-rate revenue has now surpassed $30 billion—up from approximately $9 billion at the end of 2025. - [Supporting source](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Source-stated Anthropic says the vast majority of the new compute will be sited in the United States. > The vast majority of the new compute will be sited in the United States - [Supporting source](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Source-stated Anthropic says it uses AWS Trainium, Google TPUs and NVIDIA GPUs to match workloads to suitable chips, which it says brings better performance and greater resilience for customers. > This diversity of platforms translates to better performance and greater resilience for customers who depend on Claude for critical work. - [Supporting source](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Source and contributor context Anthropic · Publisher The publisher is identified here. No individual author has been inferred. ## Source media Media retrieval status: unchecked. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Gigawatt A unit of power equal to a billion watts, roughly the output of a large power plant. - [Definition reference](https://en.wikipedia.org/wiki/Watt) ## Tensor processing unit A chip designed by Google for the calculations used in training and running AI models. - [Definition reference](https://developers.google.com/machine-learning/glossary#tensor-processing-unit-tpu) ## Graphics processing unit A chip that performs many calculations at once; the main hardware used to train and run AI models. - [Definition reference](https://en.wikipedia.org/wiki/Graphics_processing_unit) ## Anthropic Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems. - [Explore this context](/entities/anthropic) - [Reference](https://www.anthropic.com/news) - [Reference](https://www.anthropic.com/research/team/economics) ## Claude Claude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding. - [Explore this context](/entities/claude) - [Reference](https://claude.com/product/overview) - [Reference](https://www.anthropic.com/news) - [Reference](https://www.anthropic.com/research/team/economics) ## NVIDIA NVIDIA develops processors, networking and software used for AI, graphics and scientific computing. - [Explore this context](/entities/nvidia) - [Reference](https://blogs.nvidia.com/) ## Editorial provenance Version 1. Publication ID: d107f540-887a-2406-b7eb-2ceb041e2fc7. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-07T15:13:53.711Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model. Observation: 6a7c5511c444b9262a8d8989eb567f955e7d87ffc4dfe13b923149f543650116. Source text hash: cd5cbe82c19c8d42354cfa3a7d717676a8d770a0c81c2adcac089f84a73fea78. ## Revision and source-check history ### Version 1 — 2026-11-07T15:18:23.475Z Anthropic announced the agreement on April 6, 2026. It says the deal with Google and Broadcom is for multiple gigawatts of next-generation TPU capacity, expected to come online starting in 2027, to power Claude models and serve customer demand. Anthropic CFO Krishna Rao describes it as part of a “disciplined approach to scaling infrastructure” and says the company is making its “most significant compute commitment to date.” Anthropic ties the expansion to rapid growth in customer demand: it reports run-rate revenue above $30 billion, up from approximately $9 billion at the end of 2025, and says the number of business customers spending over $1 million annually has doubled since its February Series G announcement. The announcement says most of the new compute will be in the United States, extending Anthropic’s earlier commitment to invest in American computing infrastructure. Anthropic also presents its use of AWS Trainium, Google TPUs and NVIDIA GPUs, alongside access to Claude through three major cloud platforms, as a way to match workloads to hardware and provide customers with better performance and resilience. The scale of capacity that will come online, and how much it will improve service, remain unspecified. ## Original source Anthropic · Article Original source date: Apr 6, 2026 (2026-04-06T00:00:00.000Z). Date basis: Date stated on the publisher’s page. Added to Escape Meridian: 2026-11-07T15:18:23.475Z. First observed: 2026-10-19T04:45:24.603Z. Retrieved: 2026-10-28T13:53:49.638Z. - [Anthropic expands Google and Broadcom compute deal](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Limits Text as retrieved on October 28, 2026; originally published April 6, 2026. Anthropic does not specify the exact amount of capacity or how much will come online at each stage. ## Open question How much of the announced capacity will come online starting in 2027, and what effect will it have on Claude service? ## Source-stated Anthropic says its agreement with Google and Broadcom covers multiple gigawatts of next-generation TPU capacity, expected to come online starting in 2027. > We have signed a new agreement with Google and Broadcom for multiple gigawatts of next-generation TPU capacity that we expect to come online starting in 2027. - [Supporting source](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Source-stated Anthropic CFO Krishna Rao calls the expansion part of a disciplined approach to scaling infrastructure and its most significant compute commitment to date. > “We are making our most significant compute commitment to date to keep pace with our unprecedented growth.” - [Supporting source](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Source-stated Anthropic reports that its run-rate revenue surpassed $30 billion, compared with approximately $9 billion at the end of 2025. > Our run-rate revenue has now surpassed $30 billion—up from approximately $9 billion at the end of 2025. - [Supporting source](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Source-stated Anthropic says the vast majority of the new compute will be sited in the United States. > The vast majority of the new compute will be sited in the United States - [Supporting source](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Source-stated Anthropic says it uses AWS Trainium, Google TPUs and NVIDIA GPUs to match workloads to suitable chips, which it says brings better performance and greater resilience for customers. > This diversity of platforms translates to better performance and greater resilience for customers who depend on Claude for critical work. - [Supporting source](https://www.anthropic.com/news/google-broadcom-partnership-compute) ## Source and contributor context Anthropic · Publisher The publisher is identified here. No individual author has been inferred. ## Source media Media retrieval status: unchecked. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Gigawatt A unit of power equal to a billion watts, roughly the output of a large power plant. - [Definition reference](https://en.wikipedia.org/wiki/Watt) ## Tensor processing unit A chip designed by Google for the calculations used in training and running AI models. - [Definition reference](https://developers.google.com/machine-learning/glossary#tensor-processing-unit-tpu) ## Graphics processing unit A chip that performs many calculations at once; the main hardware used to train and run AI models. - [Definition reference](https://en.wikipedia.org/wiki/Graphics_processing_unit) ## Anthropic Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems. - [Explore this context](/entities/anthropic) - [Reference](https://www.anthropic.com/news) - [Reference](https://www.anthropic.com/research/team/economics) ## Claude Claude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding. - [Explore this context](/entities/claude) - [Reference](https://claude.com/product/overview) - [Reference](https://www.anthropic.com/news) - [Reference](https://www.anthropic.com/research/team/economics) ## NVIDIA NVIDIA develops processors, networking and software used for AI, graphics and scientific computing. - [Explore this context](/entities/nvidia) - [Reference](https://blogs.nvidia.com/) ## Editorial provenance Version 1. Publication ID: d107f540-887a-2406-b7eb-2ceb041e2fc7. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-07T15:13:53.711Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model. Observation: 6a7c5511c444b9262a8d8989eb567f955e7d87ffc4dfe13b923149f543650116. Source text hash: cd5cbe82c19c8d42354cfa3a7d717676a8d770a0c81c2adcac089f84a73fea78. ### uncertain — 2026-11-07T15:13:53.711Z — anthropic/claude-opus-5.5 Publication: d107f540-887a-2406-b7eb-2ceb041e2fc7 attribution — pass: This is a single Anthropic announcement on its own site. The reading attributes the deal, growth figures and hardware claims to Anthropic, and the quotes to CFO Krishna Rao, as the source does. > said Krishna Rao, CFO of Anthropic summary — pass: The reading leads with the actor and the deal, and covers the stated rationale and growth figures. Gigawatt, TPU and GPU are defined on the page. Run-rate revenue and Series G are not explained, but the growth meaning is still readable. The closing line calling the capacity scale unspecified could better say only the exact amount is unspecified, since the source does say multiple gigawatts. > Today that number exceeds 1,000, doubling in less than two months. date — pass: April 6, 2026 matches the instant originalPublished at day precision. Availability starting in 2027 is reported as Anthropic's expectation, not an outcome, and there is no hindsight framing. The page was modified in September 2026, but the reading relies only on content consistent with the April announcement. A publisher page needs no archive disclosure. > that we expect to come online starting in 2027 support — uncertain: Most claims match verbatim excerpts. However, the summary presents Claude's availability on three cloud platforms as part of how Anthropic matches workloads to hardware and delivers performance and resilience. The source ties those benefits only to training and running on diverse chips; cloud availability is a separate statement. Smallest repair: attribute performance and resilience to the chip diversity alone. The page was also edited after publication, though the cited text appears original. > which means we can match workloads to the chips best suited for them. This diversity of platforms translates to better performance and greater resilience media — pass: No media was selected and no verified candidates were supplied, so nothing is misidentified. ## Recorded references Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter. No public references recorded here yet. [Reference data (JSON)](/readings/41e47cf243a14d007caaa868758d1cc429925807ee73c4c69ca062f09369053b/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/anthropic-announces-expanded-google-and-broadcom-compute-41e47cf2.json)