This Powerhouse AI Stock Just Became More Valuable Than Broadcom (Briefly) and Meta. Here's Why It's Still an Outstanding Buy.

Taiwan Semiconductor posted a strong Q1, with revenue up 40.6% year over year, gross margin at 66.2%, and operating margin at 58.1%, while high-performance computing revenue rose 20% quarter over quarter and reached 61% of sales. Q2 guidance calls for revenue growth of 32% year over year and gross margin of 65.5%-67.5%, though management flagged 2%-3% gross margin dilution from overseas expansion in the second half and another 2%-3% hit in 2026 from 2nm ramp costs. Management said higher capex has historically preceded higher growth, supporting the bullish long-term AI demand thesis despite near-term margin pressure.

Analysis

TSMC’s read-through is less about one company and more about the persistence of the AI capex cycle. The key second-order signal is that hyperscaler demand is still outrunning foundry capacity, which means the bottleneck is shifting from “do they want to spend?” to “how quickly can the ecosystem deliver.” That favors the highest-quality capacity owners and equipment suppliers, while compressing differentiation for less efficient AI participants that need to buy the same constrained manufacturing window.

The margin guide is the important tell: near-term dilution from overseas buildout and next-gen ramps is a classic “earnings down, franchise value up” setup. Over the next 2-4 quarters, the market is likely to keep underwriting TSM’s earnings power because every incremental dollar of capex has historically supported a higher revenue base 12-24 months later. The risk is not demand destruction; it is execution slippage, especially if overseas ramp costs persist longer than expected or if leading-edge node migration introduces yield noise.

For the ecosystem, this is supportive for NVDA and AVGO because it validates that AI infrastructure spend is still absorbing supply, but the cleaner trade is often the picks-and-shovels layer that monetizes the capex wave without the same margin drag. The more interesting contrarian angle is that the market may be underestimating how much of the AI supply chain is now effectively capacity-constrained rather than demand-constrained, which can keep pricing power elevated in leading-edge tools, testing, and advanced packaging for multiple quarters.

The main reversal catalyst would be a visible pause in hyperscaler orders or a sudden step-down in AI server utilization, neither of which usually shows up first in headlines. Watch for any sign that the 2nm ramp or overseas expansions force a larger-than-expected gross margin reset into 2026; if that happens, TSM could de-rate on duration even if the longer-term growth story remains intact.

Key Decisions for Investors

• Stay long TSM on any 2-3% post-earnings pullback; use a 3-6 month horizon and treat gross-margin dilution as a better entry point rather than a thesis break. Risk/reward remains favorable as long as AI capex stays above trend.
• Add NVDA vs. a basket of AI-infrastructure laggards via a long NVDA / short SOX component pair; TSM’s results argue the strongest monetization still accrues to the most architecturally entrenched name. Timeframe: 1-2 quarters.
• Buy AVGO on dips with a 6-12 month horizon; the market is likely to keep rewarding companies that monetize AI networking and custom silicon without the same direct fab-execution burden. Best entry is on any broad semis weakness tied to margin-dilution headlines.
• Avoid shorting TSM outright; if concerned about 2026 margin pressure, express it with put spreads 9-12 months out instead of stock short exposure. That limits carry risk while preserving downside if the market starts discounting the next-node transition.
• Watch for a supply-chain long in advanced packaging / equipment as a secondary beneficiary trade over the next 1-3 quarters; TSM’s commentary implies the bottleneck is still capacity buildout, not demand. Use a basket approach rather than single-name risk.