Broadcom's Artificial Intelligence (AI) Chip Sales Surged 221% to $16.7 Billion Last Quarter: Is the Stock a Screaming Buy Right Now?

Broadcom reported AI semiconductor revenue up 221% to $16.7B in its latest quarter, alongside 86% revenue growth and 96% diluted EPS growth. For the upcoming fiscal quarter, management guided revenue to about $34.8B (+93% YoY), broadly in line with the ~$35B analyst expectation, despite a prior guidance history that has raised speculation of slight underguiding. The stock sold off after the print but the article argues the 2027 AI growth is not yet priced in, citing a forward P/E around 19x and potential for >50% upside if growth estimates are met.

Analysis

The market is still pricing AVGO like a leveraged beneficiary of AI spend, not like a durable platform with optionality into 2027. That gap matters because the real upside is less about this quarter and more about whether custom silicon becomes a second procurement standard alongside merchant GPUs; if it does, the rerating can persist for several quarters as investors extend the earnings curve.

The second-order winner is not just AVGO but its largest hyperscaler customers, especially GOOG and META, if ASICs lower inference cost per query and improve AI unit economics. The bigger loser is NVDA’s long-duration multiple if hyperscalers prove they can diversify supply without sacrificing performance; that risk is slow-moving over 6-18 months, but it can compress the premium if large customers keep signaling in-house silicon progress.

The near-term risk is that this is still a concentrated, lumpy demand story disguised as broad AI adoption. If any one of the big customers pauses capex, or if management again shades guidance conservatively and then has to lift later, the stock can remain rangebound despite strong fundamentals; the market wants evidence that growth is repeatable, not merely impressive.

Contrarian take: consensus may be underestimating how much of the upside is already in the multiple, while overestimating how easily custom silicon scales across workloads. The right falsifier is not a bad quarter, but a sequence of checks that fail to confirm the 2027 ramp: weaker guide, slower AI revenue growth, or a visible moderation in GOOG/META/Anthropic order cadence. If that happens, the stock’s earnings power is fine, but the growth-duration premium should compress.

Key Decisions for Investors

• Buy AVGO on any 3-5% post-earnings weakness and size for a 6-18 month hold; the setup is better as a compounder than a quick momentum trade. Risk/reward improves if forward guidance stays even modestly ahead of consensus, because the stock can rerate before the 2027 numbers become visible.
• Pair trade: long AVGO / short NVDA into the next hyperscaler capex cycle. The thesis is that ASIC adoption is a share-of-wallet risk to merchant GPUs over 6-18 months; invalidate it if NVDA commentary on hyperscaler demand or Blackwell supply re-acceleration remains dominant.
• Consider a barbell long GOOG + META versus broad AI infrastructure ETF exposure if you want indirect beneficiaries of lower AI inference cost, but only on confirmation that ASIC deployment is reducing unit compute costs. If capex intensity rises faster than monetization, cut the trade.
• Set a watch item on AVGO guidance quality next quarter: if revenue guide again comes in below sell-side by only a small amount, the market is likely still skeptical and the stock may stay rangebound despite headline growth. A clean beat-and-raise sequence is the catalyst that would justify expanding position size.