- Broadcom revenues are surging due to demand from hyperscalers for its AI chips
- It said for 2027 it has secured the supply to again double AI revenue to approximately $115 billion
- In 2028 it has line of sight for its AI semiconductor revenue growth to again double to $230 billion
Broadcom reported some amazing numbers for its third quarter 2026, including revenue of $29.6 billion, up 86% from the prior year period. And it gave fourth quarter 2026 revenue guidance of about $34.8 billion, an increase of 93% from the prior year period.
"Demand for our custom AI accelerators and networking continues to be very strong,” said Hock Tan, president and CEO of Broadcom. Its Q3 AI semiconductor revenue of $16.7 billion grew 221% year-over-year. For Q4, Broadcom expects its AI semiconductor revenue to accelerate to $21.7 billion, up 236% year-over-year.
Tan talked about its AI business possibly doubling next year, driven by demand for its custom accelerated processing units (XPUs).
“We are continuing to see exponential growth in demand from our XPU customers,” he said. “We believe the vast majority of compute demand for AI workloads today originates from this concentrated group who develops state-of-the-art frontier models, and we expect their need for compute infrastructure to inflect even more in 2027 and 2028. And they are all growing with XPUs to achieve superior performance, cost and power.”
Tan said, “Anthropic is on track to become our largest XPU customer in 2027 and sustain that in 2028.” Its other big customers include Google, OpenAI and Meta.
AI semiconductor revenue guidance
The company gave some general guidance for its AI semiconductor revenue in 2027 and even 2028.
It said in 2027 it has secured the supply to again double AI revenue to approximately $115 billion. In 2028 it expects the trajectory of growth to continue. “We have line of sight for fiscal 2028 AI semiconductor revenue growth to again double to $230 billion," said Tan.
Analysts on the earnings call were very interested in Broadcom’s Tomahawk 6 chips. These are high-bandwidth Ethernet switch chips designed primarily for AI and high-performance data-center networking.
Analyst Harlan Sur with J.P. Morgan, said “We've heard that Tomahawk 6 ramp has been the fastest ramp of any of your switching families, and we've also heard that you guys are almost sold out of Tomahawk 6 for next year.”
Charlie Kawwas, president of Broadcom's Semiconductor Solutions, said, “Our Tomahawk 6 has been a phenomenal success. It's available both in 100 gig and 200 gig, SerDes. So we actually have two versions of Tomahawk 6. Both are very successful, and both are deployed in pretty much all of the AI hyperscalers that are building XPUs with us. And even those that are not using our XPUs are using the Tomahawk 6, both 100 gig and 200 gig.”
Kawwas added that its customers who build XPUs with it, deploy either Tomahawk 6 or the company’s newest switch chip Tomahawk Ultra.
“We're seeing scale-up solutions adopt both Tomahawk 6 and Tomahawk Ultra,” said Kawwas. “And the beauty and the reason why they're doing this is it's Ethernet-based, which means it's open, anybody can connect to it, especially with the standards that we have collaborated with the entire industry on. So at this point in time, we're seeing it pretty much deployed in both XPU clusters and in some GPU clusters.”
Read more about Broadcom and AI on Fierce Network
Broadcom sees hyperscalers sidelining telcos for some scale across projects
Nvidia’s latest AI trick: turning storage into memory relief