Broadcom Is Down 27% From Its High. The AI Story Hasn’t Changed.

The quarter was good. Really good, actually. Revenue up 48% year-over-year. EPS of $2.44, beating estimates. AI semiconductor revenue of $10.8 billion — more than doubled on an annual basis. Free cash flow of $10.3 billion. Record operating margin of 67.3%.

And then the stock fell 14% the next day.

That’s the Broadcom situation in one sentence. When you walk in as one of the most hyped AI names in the market, good isn’t enough. Investors wanted great. What they got instead was a guidance hold — Hock Tan didn’t raise the company’s long-term AI revenue target — and a margin mix story that spooked the crowd.

What the Numbers Actually Said

Broadcom’s Q2 FY2026 results, reported June 3:

  • Revenue: $22.19 billion vs. $22.27 billion estimated — a slight miss on the top line
  • EPS: $2.44 adjusted, beating the $2.40 estimate
  • AI semiconductor revenue: $10.8 billion, up 143% year-over-year
  • Gross margin: 77.1%
  • Free cash flow: $10.3 billion
  • Q3 guidance: Revenue of approximately $29.4 billion, ahead of the roughly $28.5 billion Wall Street expected; AI semiconductor revenue guide of $16.0 billion, projecting about 200% year-over-year growth

The long-term AI revenue guidance stayed at “in excess of $100 billion” (for AI semiconductor revenue in fiscal 2027). Tan reiterated it. He didn’t push it higher. That held-flat posture is what the market punished.

Why the Selloff Happened

Expectations get dangerous when they outrun reality. AVGO had run hard heading into earnings. At those levels, the stock was pricing in a guidance raise that never came. The gross margin and mix debate added to the anxiety — even modest mix shifts can change the blended margin profile, and investors worried about where that trend goes as AI revenue becomes a larger percentage of total sales.

What’s interesting is the reaction may have been overdone. The Q3 AI semiconductor revenue guide of $16.0 billion implies about 200% year-over-year growth. The $29.4 billion Q3 total revenue guide beat consensus by close to a billion dollars. And Tan has said the company’s AI business is driven by a small set of six “core” customers — often discussed as including hyperscalers and frontier AI labs such as Google, Meta, OpenAI, and Anthropic.

The Custom Silicon Angle

This is what separates Broadcom from the rest of the chip conversation right now. While Nvidia dominates the merchant GPU market, Broadcom is quietly becoming a dominant partner for hyperscalers and AI labs building their own custom accelerators. The OpenAI Jalapeño chip, co-developed with Broadcom, is a recent example — an inference-focused processor designed around OpenAI’s large language model serving needs.

That’s not a vendor relationship. That’s a multi-year architectural integration. Switching costs are enormous. And as more players move toward custom silicon to reduce dependency on Nvidia and lower cost-per-inference, Broadcom sits directly in the path of that trend.

The company also has the AI networking side locked up. As AI clusters scale from hundreds to thousands of chips, the networking between chips becomes a bottleneck. Broadcom’s Tomahawk and Jericho product lines address exactly that. Wall Street tends to undercount this revenue stream when modeling AVGO.

Where the Stock Stands

AVGO hit an intraday high of $495.00 on June 3 — the day earnings dropped. It’s now trading closer to $360–$406, depending on the session, roughly 27% below that peak. The stock is consolidating in a support area roughly between $364 and $374.

According to 48 analysts polled by S&P Global, the consensus is Strong Buy with an average 12-month price target of $523.73 — implying more than 40% upside from recent prices. The high target sits at $650; the low at $215.

Analyst Targets (Selected)

  • Jefferies: Buy — sees “meaningful opportunity” after recent pullback
  • Wells Fargo: Overweight — raised target to $430 in January 2026
  • Macquarie: Target of $437
  • Consensus (48 analysts, S&P Global): Strong Buy, $523.73 average target

Bull / Base / Bear

Bull: Custom AI chip momentum accelerates through FY2027. The $100+ billion AI semiconductor revenue line-of-sight becomes a floor, not a ceiling. Hyperscaler relationships deepen. Margins stabilize as software revenue scales alongside AI chips. Stock reclaims $450–$500.

Base: Stock range-trades through summer as the market digests margin mix concerns. September 3 earnings become the catalyst. Q3 results either confirm the $29.4 billion guide and reignite the bull case, or margin deterioration accelerates the debate. Stock holds $370–$420.

Bear: Google shifts more TPU work to MediaTek (reported, worth watching). Custom chip customer concentration risk surfaces. Margin pressure deepens as AI mix rises. Broader tech multiple contraction pulls AVGO toward $300.

Technical Overlay

The post-earnings selloff established a support zone between $364 and $374. AVGO needs to reclaim and hold above $420 to reassert upward momentum toward its prior highs. The 50-day moving average has rolled over following the earnings gap-down. Watch the September 3 earnings as the next key inflection — Q3 results will either validate the Q3 guide or accelerate the margin debate.

What to Watch

  • Google’s next-gen TPU vendor decision — MediaTek reports suggest Broadcom may face competitive pressure on a major custom chip relationship
  • Gross margin trajectory in Q3 — whether the AI chip mix stabilizes or erodes further
  • Any upward revision to the “in excess of $100B” long-term AI semiconductor revenue target on the September 3 call
  • OpenAI Jalapeño chip commercialization timeline
  • Broader hyperscaler capex signals from Amazon, Microsoft, and Meta Q2 earnings

Bottom Line

Broadcom posted a record quarter, guided ahead of expectations, and fell hard. That’s not a story about a broken business. That’s a story about priced-in perfection meeting a guidance hold. The underlying AI demand is real. The custom silicon moat is real. Whether the stock can work from here depends on one thing: whether Q3 results on September 3 prove that the $100+ billion AI line-of-sight was a floor — or the ceiling investors feared it was.

For informational purposes only.