Analysis
Broadcom shares dropped as much as 5% in extended trading Wednesday even after the chipmaker beat Wall Street's estimates on both revenue and earnings, CNBC reported, as investors focused on guidance that came in just below expectations for the current quarter. Broadcom posted adjusted earnings of $3.32 a share on revenue of $29.59 billion, versus estimates of $3.24 and $29.36 billion -- a clean beat on both lines, with revenue up 86% year over year.
Pulse covered Broadcom's expanding enterprise footprint earlier this week as the company works through its VMware integration and Arm-related enterprise timelines; Wednesday's print is the first hard financial checkpoint since then, and it shows the AI-chip growth story is still intact -- 86% revenue growth is nowhere near a slowdown -- but the market's bar for "beat" has moved.
What's new is the guidance miss: Broadcom guided to $34.8 billion for fiscal fourth-quarter revenue against a $35.03 billion Street estimate, a gap of roughly $230 million that was enough to erase the goodwill from an otherwise strong quarter.
Broadcom has become one of the AI boom's biggest infrastructure winners by designing custom AI chips for Google, Meta, and OpenAI -- Pulse has tracked Broadcom's financing tied to Anthropic and OpenAI chip demand -- but its stock has actually lagged the broader market this year, up only about 6% in 2026 against the S&P 500's roughly 12% gain, even as revenue nearly doubled. That gap between fundamental growth and stock performance suggests investors are already pricing in a high bar for custom-silicon demand and reacting harshly to any sign the growth rate is decelerating, even marginally.
The bear read is that a $230 million guidance miss on a $35 billion quarter is noise, not signal -- but in a stock already trading at a premium multiple for AI exposure, even noise gets punished. Watch whether Broadcom's actual fourth-quarter print, due around December, comes in above or below Wednesday's guide; a beat there would suggest Wednesday's selloff was overreaction, not the start of a real deceleration.
