Broadcom shares are on the rise. What do the chipmaker's deals with Google and Anthropic mean for them?
The new agreement will provide Anthropic with computing infrastructure, and Broadcom could bring in as much as $42 billion in revenue next year, according to analysts' estimates

The partnership between Broadcom, Google and Anthropic will see the AI company access 3.5GW of computing power through Broadcom from 2027 / Photo: Poetra.RH/Shutterstock
On Monday evening, April 6, chipmaker Broadcom announced a major expansion of its AI alliance with Google and startup Anthropic, according to documents the company filed with the U.S. Securities and Exchange Commission. Broadcom has separate deals with both companies, which Anthropic, however, called a three-way "groundbreaking partnership."
As part of a long-term agreement with Google that runs through 2031, Broadcom will continue to develop custom Tensor Processing Units (TPUs), creating a direct alternative to solutions from Nvidia. Anthropic will be one of the main users of this technology: the partnership guarantees the AI startup access to 3.5 GW of capacity built precisely on these specialized chips.
The parties did not disclose financial details of the deal. Broadcom's shares are up 3.9% on the premarket on April 7 on the back of this news.
Details
The three-way partnership "between Broadcom, Google and Anthropic" announced by the AI startup's representatives suggests that Anthropic will have access to 3.5 GW of computing power based on Google and Broadcom technologies starting in 2027, MarketWatch writes. The deal "is a continuation of our disciplined approach to scaling infrastructure," Anthropic CFO Krishna Rao said. "We are building the capacity needed to serve the exponential growth of our customer base and allow Claude to define the boundaries of AI development," he added.
The agreement builds on Anthropic's commitment to invest $50 billion to strengthen computing infrastructure in the U.S., the AI company said in a statement.
Although the parties did not specify the financial component of the deal, Mizuho analysts have already assessed the potential benefits of Broadcom from cooperation with Anthropic, CNBC points out. According to strategists' forecasts, the chipmaker's revenue from Anthropic, including through this agreement, could reach $21 billion in 2026 and grow to $42 billion in 2027.
Bernstein analyst Stacey Rasgon expressed a similar opinion, MarketWatch notes. He recalled that Broadcom previously projected AI revenue of about $100 billion in fiscal 2027. "We believe this figure looks increasingly understated," he wrote, implying that the company could show more growth than investors expect.
Context
In a March financial report, Broadcom CEO Hock Tan had already talked about the scope of the partnership with Anthropic, noting the "successful 2026 launch" that provided the AI startup with 1 GW of TPU-based capacity. "In 2027, we expect this demand to grow to more than 3 GW," Tan elaborated.
Google's Tensor Processing Units (TPUs), originally created to optimize Google's own search engine, have evolved into an AI tool, Bloomberg explains. Broadcom is developing chips using Google's TPUs, offering an alternative to technology from Nvidia. According to CNBC, demand for such alternatives has recently increased due to businesses' desire to reduce their reliance on expensive Nvidia chips.
Broadcom is also working with Anthropic's main competitor, OpenAI, to build specialized chips. At the moment, both model developers rely heavily on Nvidia graphics processing units (GPUs) rented through cloud services from Amazon, Google and Microsoft, CNBC reports.
What else you need to know about Anthropic
The popularity of Anthropic has increased dramatically this year, CNBC notes. In February, the company's chatbot Claude became the most popular free app in the U.S. in the App Store. Such results Anthropic managed to achieve against the backdrop of the company's dispute with the Pentagon, as a result of which the U.S. military department recognized the AI company as "a threat to the U.S. supply chain" (Anthropic demanded from the Pentagon not to use its technology for mass surveillance of citizens, as well as in systems of fully autonomous lethal weapons).
Against this backdrop, the company said in April that user interest in its Claude chatbot was converting into financial success: Anthropic's annual revenue exceeded $30 billion (up from $9 billion at the end of last year), and the number of large corporate customers with a check of more than $1 million doubled in just two months, surpassing the 1,000-company mark.
This article was AI-translated and verified by a human editor
