Broadcom's CEO Has Line of Sight to $100 Billion in AI Chip Revenue. Is the Stock a Buy?

Broadcom CEO Hock Tan says the company has "line of sight" to generate over $100 billion in AI chip revenue in fiscal 2027. Last quarter revenue rose 29% YoY to $19.3B and management expects fiscal Q2 revenue to increase 47% YoY to $22B; the stock trades at a forward P/E of ~28 while analysts project ~41% annualized earnings growth. The outlook is supported by demand for Tomahawk 6 networking switches, locked-in leading-edge wafers and HBM supply through 2028, and visibility into six key customers (including Google, Meta, OpenAI, Anthropic), but customer concentration and data-center spending cycles are material risks.

Analysis

Broadcom’s “line of sight” narrative creates a clear winner among midstream data‑center suppliers: firms that sell networking silicon, HBM, and foundry capacity (the choke points of accelerator rollouts) will see demand growth that is both front‑loaded and lumpy. Expect order timing to be skewed toward networking and memory purchases 3–9 months ahead of accelerator installations, amplifying quarter‑to‑quarter volatility but concentrating margin capture for suppliers who pre‑sell capacity.

The biggest fragility is concentration risk interacting with capital cycle timing: a 1–2 large‑customer slowdown or a postponement of a major hyperscaler build can create a steep, fast pullback in demand within a 6–18 month window because Broadcom’s revenue is tied to multiyear build schedules rather than broad‑based transactional demand. Equally important is the structural risk that software/model efficiencies (sparsity, quantization, or centralized multi‑tenant hosting) could compress hardware intensity per model by 10–30% over several years, lowering the long‑run attach rate for XPUs and networking gear.

From a second‑order supply‑chain angle, Broadcom’s locking of wafers and HBM through 2028 de‑risks one tail but creates the opposite exposure: if the center of gravity for compute shifts to alternative architectures or on‑prem consolidation stalls, Broadcom could be long expensive capacity and face margin pressure or forced customer repricing. The market’s current valuation appears to price significant execution; the key bifurcation is whether 2026–27 deployments follow a steady ramp or a sequence of discrete customer go‑lives that compress returns into a few quarters.

Key Decisions for Investors

• Overweight AVGO (12–24 month horizon): Buy AVGO into any <10% pullback; size to 2–4% portfolio weight. Hedge with 12‑month 10–15% OTM puts (~cost 3–6% of position) to cap downside from a concentrated customer miss. Reward: 30–60% upside if multi‑customer rollouts progress; risk: limited to put cost + drawdown if hedges not used.
• Pair trade (9–18 months): Long AVGO / Short INTC, dollar‑neutral. Rationale: isolate AI infrastructure exposure vs legacy CPU cyclicality; target AVGO > INTC outperformance of 20–30%. Stop / rebalance if divergence >15% adverse.
• Tactical options protection around earnings (next 1–3 quarters): Buy AVGO 3–6 month put spreads (e.g., buy 10% OTM put, sell 20% OTM put) to protect downside at limited net cost ahead of potential timing/visibility risks. Reward: preserves long exposure to upside while capping near‑term P/L volatility.
• Event hedge (12–24 months): Maintain a small long exposure to GOOGL or META (each 1–2% weight) as a soft hedge on AI demand continuation—if hyperscalers accelerate capex, both Broadcom and these customers should rise, limiting basis risk versus a pure hardware shortfall scenario.