I Was Wrong. Broadcom Is Far From Topping Out.

Broadcom’s AI semiconductor revenue climbed from $4.4B in Q2 FY25 to $8.4B in Q1 FY26, up 106% year over year, with Q2 guidance of $10.7B and management signaling more than $100B of AI chip revenue in 2027. The company generated $8.01B of free cash flow in Q1 FY26, returned $10.9B to shareholders, and authorized another $10B buyback while raising the dividend 10%. The article is bullish on AVGO’s long-term AI demand and backlog through 2028, though it flags valuation and customer concentration risk.

Analysis

AVGO is increasingly behaving like a supply-constrained toll road on AI capex rather than a discretionary semiconductor name. The key second-order effect is that long-dated capacity commitments convert what looks like cyclical demand into something closer to contracted infrastructure revenue, which should compress the market’s skepticism around duration even if the multiple stays elevated. That is especially relevant for foundry, advanced packaging, HBM, and networking vendors: the more Broadcom locks the stack, the more scarce those upstream bottlenecks become for slower-moving competitors.

The bigger implication is competitive displacement within AI hardware. If custom silicon and networking are taking share inside hyperscaler budgets, then merchant accelerator growth can remain strong while still seeing incremental pressure on breadth of deployment and pricing power over time. NVDA remains the cleanest beneficiary of total AI spend, but AVGO has a different path: less unit volatility, more contract visibility, and a cash-return floor that should keep dip buyers engaged even if the headline growth rate decelerates.

The main risk is not demand disappearing; it is sentiment resetting on any execution miss, customer delay, or supply-chain slip because the stock is now priced for near-flawless delivery. That means the next 1-2 quarters matter more than the 3-year story, and a single negative datapoint from a major customer can still derate the whole AI complex. The market is likely underestimating how fragile consensus becomes when a few hyperscalers control the narrative, but it is also underestimating how hard it is for a competing silicon stack to dislodge a design win once the packaging, software, and networking layer are embedded.

Key Decisions for Investors

• Stay long AVGO on 1-3 month pullbacks rather than chasing strength; use weakness after any AI-cycle headline scare to build size, because the risk/reward improves when implied perfection resets before the next guidance update.
• Pair trade: long AVGO / short a basket of slower-growing AI hardware beneficiaries with less contract visibility. The thesis is that contracted demand plus capital returns should outperform names trading purely on narrative beta over the next 6-12 months.
• Maintain a core long NVDA versus AVGO only tactically hedge if the market starts rewarding visibility over growth; NVDA remains the higher-beta exposure to total AI capex, while AVGO is the lower-volatility compounder.
• Buy AVGO downside hedges around earnings or major customer events using 1-2 quarter puts, financed by selling upside calls modestly out of the money. The stock can rerate lower quickly on any sign that customer timing slips, even if the multi-year thesis is intact.
• Watch for confirmation in networking and custom silicon mix over the next two quarters; if those two engines keep compounding, add to AVGO and cut any residual short-book exposure to the AI supply chain.