Prediction: Broadcom Will Be a $3 Trillion Company by the End of 2030

Broadcom outlined AI semiconductor revenue of $58 billion in fiscal 2026, approximately $115 billion in fiscal 2027, and $230 billion in fiscal 2028, underpinning a bullish case for a $3 trillion valuation by 2030 from roughly $1.7 trillion today. Fiscal Q3 2026 revenue rose 86% year over year to $29.6 billion, AI chip revenue increased 221% to $16.7 billion, and Q4 guidance calls for $34.8 billion in total revenue including $21.7 billion from AI. The thesis assumes Broadcom can sustain roughly its current 19x forward P/E and generate about $160 billion in annual profit, though operating-margin guidance of 66%, down about 200bps sequentially, highlights mix-related margin risk.

Analysis

The investable issue is not whether AI revenue can scale, but whether Broadcom can preserve earnings conversion as custom accelerators become a larger share of mix. Custom silicon typically carries lower gross margin than proprietary merchant products and embeds more customer concentration; therefore, each point of operating-margin erosion materially reduces the earnings leverage implied by headline revenue. The market should begin valuing AVGO on incremental AI operating profit and backlog conversion, rather than revenue targets, over the next 1-3 quarters.

Broadcom's custom-chip ramp is a competitive pressure on NVDA at the margin, particularly for hyperscalers seeking lower inference cost and control over supply. It is not necessarily a zero-sum GPU replacement: successful ASIC deployment can expand total data-center power, networking, HBM, packaging and foundry demand. The more immediate second-order beneficiary is TSMC (TSM), where leading-edge wafer and advanced-packaging availability is the binding constraint; AVGO's secured capacity claims are only economically meaningful if shipments, yields and customer deployment schedules validate them.

Consensus appears too willing to annualize management's endpoint targets without applying a concentration and execution discount. A small number of buyers can defer deployments, redesign a generation, or use internal chips to gain pricing leverage; that would hit AVGO's utilization and margin simultaneously. Conversely, if quarterly AI revenue continues to outgrow guidance while operating margin stabilizes, the current valuation can support upside without multiple expansion over 6-18 months.

The key falsification signals are two consecutive quarters of AI revenue conversion below guidance, operating-margin deterioration beyond the currently indicated mix effect, or evidence that hyperscaler capex is shifting from deployment to power/network remediation. Any of these would challenge the assumption that supply availability, rather than customer absorption, is the only constraint.

Key Decisions for Investors

• Maintain or initiate AVGO as a 6-18 month core long only on post-earnings weakness or after confirmation that next-quarter AI revenue guidance is met; target a rerating driven by earnings delivery rather than a higher multiple. Exit/reduce if adjusted operating margin falls more than 300 bps from the recent run-rate without a corresponding upward revision to AI revenue.
• Express the custom-silicon substitution theme with a measured 3-6 month pair: long AVGO / short NVDA in equal beta-adjusted dollar terms, but only if AVGO reports sustained accelerator growth while NVDA data-center gross-margin guidance weakens. This is a relative-margin trade, not a call for an outright NVDA demand collapse; stop out if NVDA raises forward data-center revenue or margin guidance.
• Prefer TSM as a lower-company-specific beneficiary of sustained AVGO custom-chip volumes over the next 6-18 months. Monitor advanced-packaging capacity commentary and capex: incremental capacity additions can validate industry demand, while a sudden loosening in packaging lead times would undermine the supply-scarcity premise.
• Do not underwrite the long thesis from revenue targets alone. Set an earnings alert around disclosed customer concentration, backlog/cancellation language, AI gross-margin trajectory and conversion of supply commitments into shipped systems; absent those datapoints, treat AVGO as a watch item rather than adding aggressively.