Broadcom announced on Wednesday that it expects artificial intelligence chip revenue to surpass $100 billion by 2027, highlighting the rapidly growing demand for custom AI chips in a market largely dominated by Nvidia.
Following the announcement, Broadcom shares jumped nearly 5% in extended trading. The company also revealed a new share buyback program worth up to $10 billion, which will run through the end of the year.
Strong demand driven by Big Tech AI spending
The surge in demand for AI chips comes as major technology companies continue to expand their artificial intelligence infrastructure. Firms such as Alphabet, Microsoft, Amazon and Meta are projected to spend at least $630 billion this year on AI-related infrastructure, including chips, servers, data storage and networking systems.
Broadcom CEO Hock Tan said during the company’s earnings call that visibility into future demand has significantly improved.
He stated that Broadcom now has a clear outlook indicating that AI chip revenue could exceed $100 billion by 2027, driven by strong orders from major technology companies.
Broadcom forecasts strong revenue growth
Broadcom expects second-quarter revenue of around $22 billion, surpassing analysts’ average estimate of $20.56 billion, according to data compiled by LSEG. The company also projected AI chip revenue of $10.7 billion for the quarter.
Although Broadcom supplies semiconductors and infrastructure software, it typically does not design complete AI chips independently. Instead, the company collaborates closely with major clients such as Google, helping develop tensor processing units (TPUs), as well as working with OpenAI on custom processors designed specifically for artificial intelligence workloads.
Broadcom’s engineering teams assist customers by transforming early chip concepts into physical designs that can be manufactured by semiconductor foundries such as TSMC.
Growing AI chip orders from major clients
Industry analysts say Broadcom’s long-term revenue outlook suggests significant growth in demand. According to D.A. Davidson analyst Gil Luria, the company’s visibility into results beyond the next quarter is an indication that large customers are placing substantial orders for AI hardware.
CEO Hock Tan revealed that Broadcom is expected to deliver about one gigawatt of tensor processing units (TPUs) for AI startup Anthropic in 2026, with demand potentially rising to three gigawatts in 2027.
The company also plans to ship OpenAI’s first custom AI chip in 2027, with total deliveries exceeding one gigawatt of computing capacity.
The scale of these deals suggests Broadcom is competing with major AI chip suppliers. For comparison, Nvidia recently disclosed roughly five gigawatts of AI chip sales to OpenAI, while Advanced Micro Devices (AMD) has secured agreements worth about six gigawatts with both OpenAI and Meta.
Meta remains an important Broadcom customer
Broadcom also generates revenue through partnerships with Meta Platforms, despite reports suggesting the social media company might reduce reliance on external suppliers.
Hock Tan dismissed those concerns, stating that Meta’s custom AI accelerator roadmap, known as MTIA, remains active and that Broadcom is currently shipping chips for the project.
New chip technology and long-term growth plans
Last month, Broadcom revealed that it expects to sell at least one million chips by 2027 using its stacked chip design technology. This new product line could create an additional revenue stream potentially worth billions of dollars.
However, growth in Broadcom’s infrastructure software division slowed during the first quarter, increasing only about 1% to $6.80 billion, slightly below analysts’ expectations of $6.88 billion.
Financial results highlight strong AI momentum
Broadcom reported first-quarter revenue of $19.31 billion, representing a 29% increase year over year and slightly beating analysts’ estimates of $19.18 billion. Adjusted earnings per share reached $2.05, above the expected $2.03.
The company’s AI-related revenue more than doubled to $8.4 billion for the quarter ending February 1, driven by strong demand for custom AI accelerators and advanced networking technologies.
While Nvidia continues to dominate the AI accelerator market, Broadcom is increasingly emerging as a major competitor by offering specialized custom chips designed specifically for large technology companies building AI infrastructure.
