History Says a $1,000 Investment in Broadcom Will Be Worth This Much by 2028

Broadcom generated $16.7 billion of AI semiconductor revenue in fiscal Q3 2026, up 221% year over year, driven by custom AI-chip demand from clients including Alphabet, Meta, OpenAI, and Anthropic. The article projects AI semiconductor sales above $65 billion in FY2026, $115 billion in FY2027, and $230 billion in FY2028. Using Wall Street's FY2028 EPS estimate of $28.08 and a 30x earnings multiple, it estimates a $842 share price versus roughly $360 currently, implying 134% upside and a potential $2,340 value for a $1,000 investment.

Analysis

The investable issue is not whether custom accelerators grow, but whether AVGO captures enough value after passing through foundry, HBM and advanced-packaging costs. Hyperscaler ASIC programs are likely to create lumpy, customer-concentrated revenue with lower gross-margin quality than merchant networking or software; therefore, consensus revenue upside may not translate one-for-one into EPS upside. The key confirmation is sustained semiconductor operating-margin expansion alongside AI revenue growth, rather than design-win announcements or management TAM targets.

Custom silicon is a selective substitution threat to NVDA, not a wholesale replacement. It is most disruptive in stable, high-volume inference and internal recommendation workloads, while NVDA retains an advantage in frontier-model training, rapidly changing architectures and software portability. The second-order beneficiaries are TSMC and advanced-packaging/HBM suppliers, but capacity constraints at CoWoS and memory pricing can shift economics from AVGO to its suppliers and delay customer deployments.

Near-term sentiment is vulnerable because the implied multi-year earnings path requires several customers to simultaneously maintain extraordinary capex. Over the next 1-3 months, watch hyperscaler capex guidance and any disclosure of custom-chip deployment cadence; a single customer pushout would matter disproportionately. Over 6-18 months, the thesis is falsified if AI semiconductor growth decelerates without offsetting margin expansion, or if the custom business fails to diversify beyond a few internally funded buyers.

The contrarian view is that the market may be underpricing ASIC-induced pressure on NVDA's inference mix but overpricing AVGO's ability to monetize every unit of hyperscaler cost savings. Buyers should demand evidence that AVGO's share of the accelerator bill of materials—and not merely wafer volume—is increasing. The most attractive expression is relative rather than outright, given broad AI-capex and valuation risk.

Key Decisions for Investors

• Initiate a 6-12 month long AVGO / short NVDA beta-neutral pair only after the next earnings cycle confirms AI semiconductor growth with stable-to-higher segment operating margin. Target 15-20% relative outperformance; stop if AVGO guides AI revenue below prior run-rate expectations or NVDA demonstrates accelerating inference revenue and stable gross margin.
• Maintain a watch alert on GOOG and META quarterly capex guidance: a combined reduction or deferral of more than 10% versus consensus would be a reason to cut AVGO exposure, as customer concentration creates greater downside than the current bullish framing implies.
• For directional AVGO exposure, prefer staged equity purchases around earnings rather than chasing pre-results momentum. Add only if management discloses broader customer diversification or backlog conversion; absent this data, the risk/reward is insufficiently verifiable for an aggressive options position.
• Consider a 6-12 month long TSM / short SOXX relative position as a secondary custom-silicon expression if advanced-packaging utilization remains tight. Falsify on evidence of packaging capacity easing, a material AI-capex slowdown, or weaker-than-expected foundry utilization guidance.