Prediction: $5,000 Invested in Broadcom Could Be Worth This Much in 2028

Broadcom shares are up 2% year to date versus 12% for the S&P 500, despite strong results, with lower-than-expected guidance weighing on the stock. Management says signed supply deals support $115 billion in AI semiconductor revenue in 2027 and at least $230 billion in 2028; the article also cites $161.2 billion in Anthropic commitments for some Broadcom products, most of them non-cancellable. Under an illustrative scenario of 50% average EPS growth and a forward P/E of 17 by October 2028, a $5,000 investment today would be worth about $9,858, a 97.2% total return; the article notes these assumptions are uncertain.

Analysis

The key underwriting gap is conversion, not addressable market: supply agreements, customer commitments and management revenue ambitions do not by themselves establish when revenue is recognized, what margins it carries, or whether customers deploy the capacity on schedule. Treat the 2027–28 figures as a bull-case scenario until backlog conversion and customer funding are clearer. A shortfall in deployment could pressure both growth expectations and the multiple, especially after guidance has already disappointed.

Winners/losers: Broadcom can gain as hyperscalers diversify selected workloads from Nvidia GPUs, but custom accelerators are not a clean GPU replacement; total AI spending can support both. Marvell is a relevant custom-silicon competitor, while foundry and advanced-packaging capacity providers may capture value—or constrain shipments—if the build-out scales. Customer concentration and bargaining power could limit Broadcom’s economics even with strong demand.

Catalysts and horizon: Over 1–3 months, focus on guidance revisions, customer capex plans and evidence that contracted programs are moving into production. Over 6–18 months, shipment execution, capacity availability and realized AI margins matter more than headline commitments. A slowdown in hyperscaler investment, delayed deployments, weaker contract conversion or margin dilution would falsify the bullish case. The article’s EPS-growth and multiple assumptions are scenarios, not a reliable return forecast.

Contrarian view: The market may be discounting execution risk rather than overlooking AI demand. Conversely, treating Nvidia as a direct loser may overstate substitution: custom chips can expand workload capacity without displacing GPU demand.

Key Decisions for Investors

• Prefer a staged AVGO long on weakness rather than extrapolating the article’s two-year return math. Add only if forward guidance and reported AI revenue conversion support the contracted-demand narrative.
• Track quarterly AI revenue, segment margins, order/backlog conversion, customer capex and any disclosed shipment or capacity constraints. Reduce exposure if guidance is cut again or growth fails to convert into earnings.
• Do not short NVDA solely as an AVGO hedge: GPU and custom-chip demand may coexist. If expressing relative value, keep any AVGO-versus-NVDA pair modest and thesis-driven; reassess if custom-chip wins fail to translate into production or economics.