Here’s the thing about Broadcom. The company just posted $10.8 billion in AI semiconductor revenue for Q2 fiscal 2026 — a 143% year-over-year increase — guided Q3 AI chip revenue to more than $16 billion, and showed up with a $73 billion AI-related backlog stretching years into the future. And the stock sold off 12% in a single session.
That gap between the business and the price is worth sitting with for a moment.
What actually happened is instructive. CEO Hock Tan did not raise the full-year $100 billion AI chip sales target for fiscal 2027. That’s it. That’s the crime. Wall Street had quietly priced in a raise, the company delivered a reiteration, and sentiment collapsed faster than the fundamentals warranted. The Q2 revenue miss was $22.19 billion versus $22.27 billion estimated — less than a rounding error on a business this size. Adjusted EPS came in at $2.44 versus $2.40 expected. Net income rose 88% year over year to $9.31 billion.
The Business Doesn’t Match the Stock Price Story
Walk the numbers forward. Q1 AI revenue was $8.4 billion, up 106% year over year. Q2 came in at $10.8 billion. Guidance for Q3 is more than $16 billion. The sequential acceleration isn’t stalling — it’s compressing into a steeper slope. Full-year fiscal 2026 consensus revenue estimates have been revised upward to approximately $94.7 billion, with EPS estimates tracking toward $6.81 per share.
Broadcom has delivered returns of roughly 745% over the past five years, compounding through every skeptical cycle the market has thrown at it. The 200-day moving average remains intact. The stock’s all-time high of $481.57 was reached in June 2026, just weeks before this earnings-driven reset to approximately $360.
Slight tangent, but it matters: Broadcom isn’t a pure-play semiconductor story, and that’s often what the market misses. The VMware acquisition turned this into something structurally different — a company that earns infrastructure software fees that compound through the cycle regardless of quarter-to-quarter chip order timing. The software segment contributes meaningfully to the adjusted EBITDA margin, which was guided at 68% for Q2. You don’t get 68% EBITDA margins from silicon alone.
What Sophisticated Capital Is Actually Doing
The custom AI chip franchise is the part analysts should be focused on. Broadcom is deepening multi-year chip deals with major hyperscalers including Google, Anthropic, OpenAI, and Meta, while co-developing a $35 billion AI infrastructure financing platform alongside Apollo and Blackstone. The TPU-based compute agreements for Anthropic alone represent more than one gigawatt of capacity in 2026 and another five gigawatts starting in 2027. These are not one-quarter relationships.
S&P Global’s survey of 48 analysts shows a strong buy consensus, with an average price target near $524 — implying roughly 45% upside from the current price of approximately $360. The 26-analyst consensus at Public.com shows a similar Buy rating with targets clustered near $500.
The bear case is real, though. Margin pressure tied to lower-margin custom AI accelerators is showing up in the numbers. Customer concentration — with a handful of hyperscalers representing the bulk of AI revenue — creates binary risk if any major relationship shifts. And insider activity has been mildly negative, with a senior executive selling approximately 25,000 shares near $387 in late June.
Technical Structure
The stock pulled back from its all-time high of $481.57 to the current range near $360, a roughly 25% drawdown that placed it squarely on the 200-day moving average — a level that has historically served as a reentry point during prior Broadcom correction cycles. Volume on the earnings-day selloff was elevated but has since normalized. The longer-term trend structure remains bullish as long as the $320–$330 support zone holds.
VWAP from the Q2 earnings day sits near $390. Bulls need a weekly close back above that level to confirm that the post-earnings reset is complete. A failure to reclaim $390 over the next several weeks would extend the consolidation and potentially test the $320 zone where longer-term institutional accumulation patterns have historically emerged.
Three Scenarios Into Q3 Earnings
Bull Case: Q3 AI chip revenue arrives above $16 billion guidance, Hock Tan raises the $100 billion fiscal 2027 target, and margin pressure proves transitory as the revenue mix improves toward higher-margin networking solutions. Stock reclaims $450 and challenges the all-time high.
Base Case: Q3 delivers at the high end of guidance, the $100 billion target is reaffirmed without a raise, and the stock stabilizes in the $380–$420 range as institutional buyers absorb the earnings discount. Full-year revenue tracks to approximately $94 billion with EPS near $6.80.
Bear Case: A hyperscaler signals reduced custom chip orders, margin compression deepens, and the stock tests the $300–$320 support zone. This scenario would likely require a broader AI capex pullback, which current hyperscaler guidance does not support.
Active Trader Strategy Framework
- Key level to reclaim on the upside: $390 VWAP from the earnings day selloff
- Primary support zone: $320–$330, where longer-term institutional buyers have historically stepped in
- Position sizing should account for elevated Q3 earnings volatility — options implied volatility tends to reprice sharply heading into the late-August report date
- Risk management consideration: the bear case is a genuine $300 test if the broader AI capex theme softens, which warrants defined risk frameworks over large directional positions
- Watch the custom silicon signals — any hyperscaler commentary on XPU versus GPU allocation shifts materially impacts the Broadcom thesis
The market sold a 143% AI revenue growth quarter because management didn’t exceed its own target. That’s the mispricing. The $73 billion backlog doesn’t evaporate because one earnings call underwhelmed. What happens at the Q3 report — whether guidance is raised, maintained, or cut — is the single most important catalyst between now and the end of the year.
For informational and educational purposes only. Not investment advice. Trading involves risk, including loss of principal.