Broadcom Stock Is Back Where It Ended 2025, but Its Earnings Are About 43% Higher. Is It a Buy?

Broadcom raised its AI-chip revenue outlook to about $58B for fiscal 2026, $115B for fiscal 2027 and $230B for fiscal 2028, while targeting more than $30 in fiscal-2028 adjusted EPS. Fiscal Q3 AI-chip revenue rose 221% year over year to $16.7B, and adjusted EPS increased 96%, but the stock remains roughly 29% below its $495 52-week high at about $350. The key risk is rising customer concentration: the top five customers generated about 55% of Q3 revenue, while Broadcom may backstop up to $29B in lease payments for one AI customer.

Analysis

AVGO's multiple discount is better understood as a change in earnings quality than a simple valuation anomaly. The incremental AI revenue is increasingly tied to a small set of hyperscaler/lab counterparties, including customers requiring structured financing; this converts part of the opportunity from semiconductor execution risk into credit, residual-value, and project-finance risk. If deployment schedules slip, AVGO could face both revenue deferrals and impairment/guarantee exposure, while a conventional P/E framework will understate the downside.

Near term (days to 3 months), raised backlog-supported targets should support AVGO on dips, particularly if quarterly AI revenue continues to exceed the implied exit run-rate. The more important catalyst is disclosure around financing commitments, receivables, lease backstops, and customer-specific concentration at the next earnings call; any increase in guarantees without commensurate cash collections should pressure the multiple even if revenue guidance holds. Over 6-18 months, custom ASIC adoption is a real competitive threat to NVDA's accelerator share, but it is not a clean zero-sum outcome: higher total cluster buildouts also sustain NVDA networking, memory, foundry, and power/cooling demand.

Consensus appears too focused on whether the nominal 2028 revenue target is attainable and too little on who funds it. The contrarian bear case is that AVGO is effectively subsidizing AI-lab capacity to secure design wins, making reported growth less valuable than demand funded from hyperscaler balance sheets. Conversely, if financed customers demonstrate third-party funding and rising utilization, AVGO can re-rate because the market is currently assigning a material probability to delayed conversion rather than permanent demand loss.

Key Decisions for Investors

• Accumulate AVGO in tranches over the next 1-3 months rather than chase a single entry; target a 12-18 month long only if quarterly AI revenue and cash conversion remain aligned. Risk/reward is favorable if the multiple normalizes, but cap exposure until financing disclosures are independently verified.
• Use a defined-risk AVGO call spread dated 9-12 months out, funded only after the next earnings release confirms no material expansion in lease-payment or residual-value guarantees. Thesis invalidation: reduced fiscal-2027 AI outlook, weaker cash collections, or incremental customer-financing commitments disproportionate to revenue.
• For a relative-value expression, long AVGO / short a basket proxy of AI infrastructure names with the most stretched revenue assumptions is preferable to an outright NVDA short; avoid a direct AVGO-vs-NVDA pair because custom ASIC growth can expand total AI capex and benefit both companies.
• Monitor APO and BX as second-order credit-risk beneficiaries, not core longs: expanded AI infrastructure financing can generate fee-bearing AUM, but losses or mark-downs on concentrated AI-lab exposures would damage fundraising and realization narratives. Treat disclosures of actual funded commitments, not announced facilities, as the trading catalyst.