2 Companies That Will Join Nvidia, Apple, and Alphabet in the $3 Trillion Club by 2028

Broadcom ($1.59T) and TSMC ($1.75T) are positioned to benefit from AI infrastructure build-outs; Broadcom targets $100B in AI chip revenue by end-2027 and reported $8.4B in AI semiconductor revenue in Q1 FY2026. Analysts peg Broadcom at $17.54 EPS for 2027 and the piece models a 40x P/E implying $701/share (≈109% upside); TSMC projects ~25% CAGR 2024–2029, with a modeled 95% stock rise to reach $3T by 2028. Recommendation tone is bullish for both names, with potential single-stock moves but limited immediate market-wide impact.

Analysis

Broadcom’s co‑design/custom-AI strategy creates a different durability profile than commodity GPU suppliers: revenue becomes sticky via integration, firmware/IP handoffs and multi‑year validation cycles, which raises switching costs and supports sustained gross‑margin expansion if Broadcom secures a handful of large hyperscaler anchors. That stickiness also concentrates counterparty risk — a lost anchor or an in‑house silicon pivot by one hyperscaler would compress forward booked revenue quickly; conversely, incremental wins compound because each new design increases tooling/licensing leverage across suites of network/storage silicon.

TSMC’s neutral-foundry position gives it optionality to capture demand across architectures, but the real lever is capacity cadence: multi‑year lead times on EUV tools and wafer starts mean that cyclical spending now locks revenue mix out several years, creating a multi‑year supply rhythm that benefits tool vendors and raises the floor on pricing for advanced nodes. The flip side: geopolitical disruption in the Taiwan Strait or tighter export controls would create an asymmetric immediate premium to insured capacity (and a correlated spike in tooling/order cancellations) that could rerate not just TSMC but the whole advanced toolchain.

Second‑order winners include advanced toolmakers and cloud customers that can monetize marginal efficiency gains (fewer GPUs per workload) while losers include commodity GPU incumbents if custom ASICs take share for steady-state inference workloads. Key near‑term catalysts are confirmed multi‑year design wins, wafer‑start cadence disclosures, and major tool deliveries; watch those as binary re‑rating events over 3–18 months.

Key Decisions for Investors

• Long AVGO via defined-risk call spread (12–18 month): buy 1 ATM call / sell a higher strike to finance ~60–80% of premium. Position size 3–5% notional. Rationale: capture re‑rating from co‑design wins with capped capital at risk; target 40–80% net return if Broadcom converts several hyperscaler design wins within 12 months, max loss = premium (~3–5% of notional).
• Pair trade — Long TSM (18–24 month) / Short INTC (18–24 month) equal notional exposure: overweight foundry secular share vs IDM execution risk. Use 10% of position to buy 1‑year 15% OTM puts on TSM as geopolitical insurance. R/R: asymmetric — expect 40–70% upside on TSM if capacity ramps go to plan, while INTC downside hedges macro and node‑execution risk.
• Tail hedge: buy 12–24 month out‑of‑the‑money puts on TSM (10–20% OTM) sized to cover 20–30% of total Asia production exposure. Cost is insurance against Strait/geopolitical closure which would produce >2x volatility in foundry multiples.
• Event‑driven rule: add incremental AVGO only after confirmed >$1B multi‑year customer disclosure or equivalent design-win milestone; trim 25–40% of position on a +40% price move or if bookings miss. This keeps exposure aligned to realised bookings rather than hype.