The chipmaker's AI semiconductor business is now on pace to more than triple year-over-year, with quarterly revenue projected at $29.4 billion.
Broadcom just posted a quarter that makes most of the semiconductor industry look like it’s standing still. The company’s Q3 fiscal year 2026 revenue came in at roughly $29.4 billion, an 84% jump compared to the same period last year, with AI semiconductor revenue alone projected above $16 billion.
That AI figure represents growth north of 200% year-over-year. To put that in perspective, Broadcom’s entire AI chip business in one quarter is now larger than what many standalone semiconductor companies generate in a full year.
AI is now the majority of the business
The most striking detail isn’t just the raw revenue number. It’s the composition. AI semiconductors are expected to account for roughly 54% of Broadcom’s total quarterly revenue, meaning the company’s identity has fundamentally shifted.
The trajectory has been steep. In Q2 FY2026, Broadcom reported actual revenue of $22.19 billion, up 48% year-over-year, with AI semiconductor revenue of $10.8 billion, a 143% increase. Going from $10.8 billion to above $16 billion in a single quarter is the kind of sequential acceleration that typically makes analysts re-examine their models.
AI networking contributed nearly 40% of AI revenue in Q2, highlighting that Broadcom’s opportunity extends well beyond custom silicon.
CEO Hock Tan and CFO Kirsten Spears have positioned AI semiconductors as the company’s primary growth engine.
The client list explains the confidence
Broadcom’s AI business isn’t built on speculative demand from unnamed customers. Its major clients include Google, where Broadcom designs multiple generations of the Tensor Processing Unit (TPU), along with Meta, Anthropic, and OpenAI.
AI bookings in Q2 exceeded $30 billion, which provides substantial visibility into future quarters.
For the full fiscal year 2026, Broadcom raised its AI semiconductor guidance to $56 billion, representing approximately 180% growth from fiscal year 2025. And the company has already indicated it expects AI semiconductor revenue to exceed $100 billion in FY2027.
Margins tell the profitability story
Revenue growth at this scale is impressive, but it only matters if it’s profitable. Broadcom’s non-GAAP operating margin is anticipated to stay around 67% for Q3, a figure that would make most software companies jealous, let alone hardware manufacturers.
Maintaining margins at that level while growing revenue 84% year-over-year suggests Broadcom has significant pricing power and operational efficiency in its AI business. Custom silicon designed for hyperscalers tends to command premium pricing because switching costs are enormous. Once Google builds its infrastructure around Broadcom-designed TPUs, migrating to a competitor becomes a multi-year, multi-billion-dollar endeavor.
What this means for the competitive landscape
For investors tracking the AI semiconductor space, the concentration of revenue among a small number of hyperscale customers remains both a strength and a risk. Google, Meta, Anthropic, and OpenAI are spending at historically unprecedented levels on AI infrastructure.
The $100 billion FY2027 AI semiconductor revenue target is the number worth watching. Hitting it would require Broadcom to nearly double its AI business again in a single fiscal year.
Still, the Q2 bookings figure of over $30 billion provides a tangible floor. Customers don’t commit that kind of capital on a whim. Those orders represent physical data centers being planned, power infrastructure being built, and deployment timelines that stretch years into the future.
