Broadcom says that big AI companies and hyperscalers can’t develop and produce their own chips at scale for now. According to the company, this is clear from the strong demand for custom chips that Broadcom designs for Meta, OpenAI, Anthropic, and others.
During the presentation of the first-quarter fiscal year 2026 quarterly figures, CEO Hock Tan reported that revenue from AI-related chips grew 106 percent year over year to $8.4 billion.
Broadcom is currently developing custom AI accelerators for five major customers, reports The Register. The company expects demand from Google to increase further when a new generation of TPUs is rolled out. Other AI companies are also preparing large-scale implementations. Anthropic plans to deploy approximately one gigawatt of TPU capacity developed with Broadcom technology in the near future and is already looking ahead to an expansion to three gigawatts in 2027.
Meta plans to deploy several gigawatts of computing power based on Broadcom’s XPU accelerators during the same period. OpenAI is also said to plan to run more than 1 gigawatt of compute on such chips by 2027.
According to Broadcom, the company has now secured the supply chain to meet this demand until at least 2028, including critical components such as high-bandwidth memory. Tan expects Broadcom to be able to sign similar contracts in the coming years, as many technology companies are finding it difficult to build a complete chip strategy on their own.
According to the CEO, designing a chip is only part of the challenge. Companies also need to attract specialized designers, organize manufacturing processes, master packaging techniques, and efficiently connect chips in large-scale systems. In addition, their own chips must compete with solutions from established suppliers such as Nvidia and other AI platform providers.
According to Tan, it is relatively easy to design a chip that works in a laboratory. The real challenge lies in mass-producing hundreds of thousands of units at an acceptable cost and with sufficient yield in the production process.
Strong growth in AI network hardware
Broadcom’s network division is also benefiting from AI growth. Revenue in this segment rose by 60 percent year-on-year. The company is currently working on a new generation of Tomahawk switch chips, due to be released next year and expected to double the performance of the current model.
Total semiconductor revenue for the past quarter was $12.5 billion, an increase of 53 percent. Revenue from chips unrelated to AI remained stable at $4.1 billion.
Broadcom’s software division, which includes CA, Symantec Enterprise, and VMware, saw limited growth. Revenue came in at $6.8 billion, up 1%. VMware grew by 13 percent and, according to Tan, is benefiting from the growing demand for enterprise AI infrastructure.
For the second quarter, Broadcom expects total revenue of approximately $22 billion. Investors responded positively to the figures and the announcement of a new share buyback program. Broadcom’s share price rose by almost five percent in after-hours trading.