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

Broadcom said AI semiconductor revenue surged 221% to $16.7B in the latest fiscal quarter, alongside a total company revenue gain of 86% and 96% growth in diluted EPS. For the next quarter (fiscal Q4), management guided revenue up ~93% YoY to $34.8B, slightly above the ~$35B level analysts expected, after prior guidance of $29.4B for the earlier period. Despite a post-earnings sell-off and a lackluster 2026 stock performance, the article argues the custom AI chip momentum (ASICS for hyperscalers including Alphabet, OpenAI, Anthropic, and Meta) suggests upside if 2027 growth estimates play out.

Analysis

AVGO looks like the cleanest public-market expression of the hyperscaler move from brute-force GPU buying to more capital-efficient custom silicon. That matters because it changes who captures the economics: vendors like AVGO gain sticky design-win annuities, but the bigger margin uplift can accrue to GOOG/GOOGL and META if custom accelerators lower inference cost per token and free cash for additional AI spend. NVDA is not an obvious short on absolute demand, but the mix effect is real: every incremental workload migrated to ASICs is one less dollar of merchant GPU content, which can cap multiple expansion in the semis group.

The near-term trade is about credibility, not the long-term AI story. Investors are discounting the 2027 bridge because custom-chip revenue is lumpy, concentrated in a handful of customers, and easy to overextrapolate from one strong print; the stock can stay range-bound unless management converts this into a clearer multi-quarter cadence. The key falsifier is any sequential deceleration or a guide that fails to raise conviction around the out-year earnings path; if that happens, the rerating narrative breaks and AVGO trades more like a mature semi than an AI compounder.

Contrarian view: the market may be underestimating the second-order winner set. If hyperscalers can source more of their own silicon, the better long may be GOOG/GOOGL or META rather than AVGO, because they get the cost savings without paying the scarcity multiple that semis usually command. So the consensus is probably right that AVGO is a quality AI beneficiary, but wrong to assume the upside is purely linear from here; the best risk/reward may be relative-value rather than outright beta chasing.

Key Decisions for Investors

• Buy AVGO on 3-5% post-print weakness over the next 1-2 sessions; target a 3-6 month rerating if management can keep AI growth compounding, with downside if the next guide or commentary fails to support the 2027 bridge.
• Pair trade: long AVGO / short NVDA for a 3-6 month relative-value expression of custom ASIC share gains versus merchant-GPU mix pressure; keep sizing modest because both names still benefit from AI capex growth.
• Prefer long GOOG or META versus a broad semis basket (SOXX) for a 6-12 month horizon if the thesis is that custom silicon improves hyperscaler economics more than it boosts supplier economics; invalidate if cloud capex growth slows or custom-chip adoption stalls.