Broadcom's Latest Prediction Makes the Stock a No-Brainer Buy Right Now

Broadcom forecasts AI semiconductor revenue of $230 billion in fiscal 2028, doubling its newly raised $115 billion fiscal 2027 target; Q3 fiscal 2026 AI semiconductor revenue was $16.7 billion, or roughly $67 billion annualized. Management said it has secured the supply chain needed to support the expansion, with Alphabet, Meta, Anthropic, and OpenAI increasing spending on Broadcom-designed custom AI chips. The article argues that AVGO's 19x forward P/E does not fully reflect its projected AI growth trajectory.

Analysis

The key underwriting issue is not valuation but concentration and capex circularity: a $230B custom-silicon run rate would require a small set of hyperscalers to sustain sharply rising AI infrastructure budgets while converting inference workloads from merchant GPUs to internally designed accelerators. That shifts bargaining power toward AVGO if its IP, networking attach, and packaging access remain scarce, but it also gives each customer material leverage over pricing and creates a binary risk around any delayed tape-out or deployment pause. The revenue target should therefore be treated as a capacity-and-design-win indicator, not as booked revenue.

Near term, AVGO can outperform NVDA if investors assign a higher probability to ASIC share gains in inference, where cost per token and power efficiency matter more than software flexibility. The second-order beneficiaries are TSMC and advanced-packaging suppliers, while merchant GPU economics face a longer-dated mix risk; the relevant question is whether custom silicon expands total AI compute demand or displaces GPU purchases. GOOG and META may receive a modest multiple benefit if lower unit compute costs improve AI monetization margins, although their absolute capex burden remains the more important stock driver.

The consensus appears too willing to annualize an aspirational endpoint while underweighting customer concentration, supply-chain execution, and the likelihood that accelerator pricing compresses as volumes scale. Over the next 1-3 months, AVGO needs corroboration through backlog, customer capex commentary, and gross-margin durability; over 6-18 months, the thesis is falsified by a material cut in hyperscaler AI capex, a missed custom-chip production ramp, or AI semiconductor growth that requires margin concessions. A sustained narrowing of AVGO's growth premium versus NVDA despite rising AI revenue would signal the market is discounting these risks rather than missing them.

Key Decisions for Investors

• Maintain an overweight AVGO versus NVDA through the next two earnings cycles, sized as a pair rather than an outright beta trade. The catalyst is evidence that custom silicon is taking incremental inference share; reassess if AVGO's AI growth guide decelerates materially or gross margin falls despite volume growth.
• Use post-earnings pullbacks, rather than pre-guidance momentum, to add AVGO exposure. Require confirmation of customer-funded demand and supply availability in the earnings release; without disclosed backlog, customer concentration, or margin data, treat the long as a watch item rather than underwriting the fiscal-2028 target.
• Express the supply-chain read-through selectively via long TSM versus a semiconductor-equipment basket only after foundry utilization and advanced-packaging lead times tighten. This captures custom-chip volume while avoiding direct exposure to AVGO's customer-concentration risk; exit if hyperscaler capex guidance weakens.
• Monitor GOOG and META capex guidance as the principal falsification trigger for AVGO. A broad-based reduction in AI infrastructure spending would likely hurt AVGO's multiple before reported revenue, making AVGO downside protection appropriate around hyperscaler earnings if the position is above benchmark weight.