MarketsAVGOSEP 03, 2026

AVGO Q3: $29.6B print, $16.7B AI chip revenue, and a $230B FY2028 roadmap

Broadcom guided Q4 to $34.8B — light of the $35.03B consensus — but Hock Tan's confirmed AI ramp to $115B in FY2027 and $230B in FY2028 is the number that resets small-business AI budgeting.

BY HENLEY MARRAST · MARKETS DESK

Broadcom printed $29.59 billion in fiscal Q3 revenue, up 86% year-over-year, with AI semiconductor revenue tripling to $16.7 billion (221% YoY, 54% sequentially). Non-GAAP diluted EPS came in at $3.32 against the $3.24 LSEG consensus, a 96% year-over-year gain. Free cash flow reached $13.67 billion, up 95%. And the stock fell more than 3% in extended trading.

That last fact is the analytically interesting one, because it tells you what the market was actually pricing. Q4 guidance of roughly $34.8 billion, implying 93% growth, came in under the $35.03 billion LSEG consensus. On a print this large, a $230 million shortfall is a rounding error. The market treated it as a signal anyway.

The real news is further down the release. CEO Hock Tan guided full-year FY2026 AI semiconductor revenue to $58 billion, then walked the number forward: $115 billion in FY2027, $230 billion in FY2028. Broadcom has line of sight to double AI revenue to $115 billion in FY2027 and double again to $230 billion in FY2028, with supply already secured and demand running ahead of the current 2027 outlook.

The customer roster is the reason to believe him. Tan expects to ship “tens of billions of dollars” of processors to Google annually for several years. Broadcom is looking for Anthropic to deploy 5 gigawatts of TPU 8i chips in 2027, with line of sight to another 10 gigawatts. OpenAI’s Jalapeno chip is heading toward a 1.3-gigawatt deployment in 2027, with a second generation already taping out. Meta rounds out the confirmed hyperscaler book.

The implication for anyone buying AI-powered software downstream is uncomfortable. The infrastructure layer under every AI sales and marketing tool isn’t cheapening on the curve buyers have been assuming. It’s scaling into a $230 billion annual chip line by FY2028, and the application-layer pricing built on top has to clear that cost. The AI earnings gauntlet framing from last week called this the price signal every small business should read; the Marvell $2.74B print and Google’s $120B custom-chip commitment said the same thing from the other side of the trade.

The market sold a $230 billion roadmap because the next quarter missed by $230 million. That’s the tell.

Sources

About the author

Henley Marrast

MARKETS DESK

Henley Marrast covers AI-equity flow, accelerator demand, and earnings prints for AI Sheet Report. She leads coverage of the public AI complex from the New York markets desk, with a focus on the daily tape and quarterly results. She has been writing about technology markets for several years.