Broadcom Earnings: The Un-Raised Number That Sank AVGO

Broadcom's Q2 FY26 earnings looked like a monster—AI semiconductor revenue doubled to $10.8 billion, up 143%, total revenue hit $22.2 billion, up 48%, and Q3 guidance calls for roughly $29.4 billion, up 84%. The stock still cratered about 13%. The crowd calling that an overreaction is wrong. The tell isn't in the beat—it's in what management refused to do: raise its full-year AI target. At Broadcom's valuation, "in-line and reaffirmed" is a de-rating event, and the market priced it correctly. 

Every headline screams the same thing: AI revenue doubled, so why did AVGO fall? Wrong question. The right one is why a company supposedly in a vertical AI ramp—one that just beat its own $10.7 billion AI guide—wouldn't lift its annual number by a single dollar. The answer is the whole story.

The numbers the bulls are quoting:

• Revenue $22.2B (+48%); non-GAAP EPS $2.44, ahead of the ~$2.40 consensus.
• AI semiconductor revenue $10.8B (+143%), roughly half of total sales. 
• Record adjusted EBITDA of $15.2B (69% of revenue) and $10.3B free cash flow. 

• Q3 AI revenue guided to about $16B, up over 200%; six core custom customers including Anthropic, Google, Meta, and OpenAI. 

The numbers the bulls are ignoring:

• Full-year FY26 AI revenue was reaffirmed at $56 billion—not raised—even after Q2 beat plan. Companies riding genuine upside raise the year. Broadcom didn't. 
• Revenue actually came in slightly below expectations; the "beat" was carried by EPS and margins, not the top line. 
• Tan said Broadcom will now sell "chips only" rather than the complete integrated AI systems it had previously promised customers—a quiet haircut to dollar content per deployment, dressed up as focus. 
• The Q3 AI guide of $16B reportedly landed under the loftiest buy-side whispers (~$17B), and the much-cited FY27 ">$100 billion" target was reiterated, not raised—a back-end-loaded number still 18 months out. 

Here's the one-line meaning: the second derivative flattened. The rate at which Broadcom raises its AI guidance just went to zero, and a stock priced for relentless upward revisions cannot tolerate "as expected." A ~13% drop—CNBC pegged the slide near 15%, closing near $419 on roughly double-average volume—isn't panic. It's repricing. 

Background: Broadcom (AVGO) supplies custom AI accelerators and AI networking chips, the picks-and-shovels of the hyperscaler buildout, alongside infrastructure software. Q2 ended May 3, 2026; results landed after the close on June 3. The AI line has carried the entire growth story—and therefore the entire valuation.

FAQ — Was the AVGO drop an overreaction? No. The bar wasn't "double AI revenue"; the market already knew that was coming. The bar was a guidance raise that signaled acceleration beyond plan. Management declined to clear it, and a perfection-priced stock de-rated accordingly.

FAQ — But isn't doubling AI revenue bullish? It's bullish for the business, not necessarily for the stock at this price. Growth that's already in consensus isn't a catalyst; only surprises move richly valued names, and the only surprise here was a withheld raise.

Forward look / risk: The bull rebuttal is real—if the $56B target was deliberately conservative and gets raised next quarter on lumpy, multi-gigawatt hyperscaler deployments, today's sell-off ages badly. But the burden of proof has flipped: until Broadcom actually raises the number, the smart read is that the easy upside surprises are behind it, and "chips only" caps how much each AI deal can ultimately bill. Buying this dip is a bet that management was sandbagging. The tape says it wasn't.

Trade US stocks/Hong Kong stocks/Taiwan stocks ,etc Online Global Trending Stocks

Start Trading Now >