Why we're raising our price target on Broadcom despite its post-earnings sell-off

Broadcom posted solid fiscal Q2 results, with revenue of $22.19B versus $22.27B consensus, adjusted EPS of $2.44 versus $2.40 expected, and adjusted EBITDA of $15.24B versus $15.06B. AI semiconductor revenue jumped 143% year over year to $10.8B, and management reiterated FY2026 AI revenue of $56B and FY2027 at least $100B while guiding Q3 revenue to about $29.4B above estimates. Despite the positive outlook, shares fell after hours because investors wanted a bigger upside surprise, especially in AI revenue, and the stock has already rallied more than 80% over the past year.

Analysis

AVGO is still the cleanest “picks-and-shovels” way to express the AI capex cycle, but the market is starting to treat it like a duration asset: the stock now needs visible near-term upside to sustain a multiple that already discounts a very large share of 2027 growth. The key second-order read is that the financing SPV effectively reduces customer budget friction, which should extend the runway for custom silicon adoption and lower the probability of a 2026 demand air pocket. That matters more than the quarter itself because it turns what had been a capex-constrained narrative into a credit-enabled deployment cycle.

The competitive takeaway is less about Broadcom vs. Nvidia and more about “cloud house” diversification. If Google, Meta, OpenAI, and Anthropic keep pushing custom silicon, the winners are network attach vendors, advanced packaging, HBM ecosystem names, and alternative lenders exposed to AI infrastructure financing; the loser is the assumption that one GPU platform monopolizes inference economics. Tan’s acknowledgment that some designs can be lost is important: it implies the custom silicon market is becoming multi-sourced earlier than bulls expected, which should cap the long-term share gains of any single OEM while still leaving aggregate demand intact.

The softer software print is not just a mix issue; it signals that investors are now valuing AVGO primarily on AI optionality, not on the slower software annuity. That makes the stock more vulnerable to any hint that AI bookings are being pulled forward rather than converting into revenue on schedule. Near term, the biggest risk is not a demand miss but a valuation reset if the next catalyst is simply confirmation rather than acceleration.

The contrarian view is that the selloff may be too harsh because the order book-to-revenue gap suggests backlog visibility remains exceptional. If management is right on 2028 growth, the market is underpricing the compounding effect of a larger installed base plus networking attach, especially into 2H26 and 2027. The tradeoff is that the stock likely needs one more step-up in guidance or a fresh customer win before it can re-rate materially higher from here.

Key Decisions for Investors

• Maintain a core long AVGO, but only on pullbacks over the next 1-3 weeks; risk/reward improves if the market de-risks the multiple while the 2027 visibility stays intact.
• Use AVGO call spreads into the next 1-2 quarters rather than outright stock to express upside from AI revenue re-acceleration; capped risk fits a name where the issue is valuation, not thesis.
• Pair trade: long AVGO / short a higher-beta AI hardware basket over 1-3 months if you expect custom silicon to keep compounding while more speculative AI names remain dependent on fresh guidance beats.
• Add BX on weakness as a financing-enabler hedge to the AI infrastructure complex over 3-6 months; the SPV structure suggests private credit/alt-finance participation is becoming a real monetizable layer in AI capex.
• Trim or hedge AVGO if the stock reclaims its pre-earnings peak without a new customer or a raised AI guide; upside likely needs a new catalyst, not just confirmation.