Taiwan Semiconductor CEO just dropped a hint about the next move in AI stocks

TSMC said AI-related demand remained "extremely robust" and lifted its full-year revenue growth forecast to over 30%, signaling continued strength in the AI supply chain. First-quarter net profit rose 58.3% year over year to NT$572.5 billion ($18.15 billion), while revenue increased 35% to $35.9 billion and gross margin expanded to 66.2%. The company also indicated it may lean toward the high end of its $56 billion capex budget to support AI infrastructure buildouts.

Analysis

The key second-order read-through is that AI capex is no longer just a GPU story; it is becoming a full-stack capacity expansion story anchored by leading-edge foundry throughput. If TSM is leaning toward the high end of its capex plan while preserving gross margin, that implies pricing power is still ahead of supply normalization, which is usually the most bullish setup for downstream accelerants like advanced packaging, HBM, EDA, and networking silicon. In that regime, the market typically over-weights near-term chip demand and under-weights the multi-quarter bottleneck relief that supports revenue revisions across the ecosystem.

The biggest beneficiary remains NVDA, but the more interesting setup is AVGO because agentic AI increases inference intensity, custom silicon, and network fabric demand at the same time. A shift from “queries” to “actions” raises token consumption per workflow, which expands not only accelerator demand but also switching, optics, and load-balancing requirements in data centers. That creates a broader basket effect: AI infrastructure spending becomes less correlated to a single model release and more tied to enterprise workflow adoption over the next 6-18 months.

Risks are mainly timing and digestion. The stock reaction suggests the market already believes in AI strength, so the near-term hazard is valuation compression if capex news becomes confirmation rather than surprise, or if customer concentration turns into order lumpiness. The other watch item is whether TSM’s capex intensity pressures free cash flow in 2H26; if margins hold while spending rises, the signal is positive, but if utilization softens before demand broadens, the group can de-rate fast.

The contrarian point is that this is a supply discipline signal masquerading as a demand signal: if TSM is committed to more capex, it may be more confident about long-duration demand than the street, but it also means the industry is still several quarters away from true supply relief. That favors owning the infrastructure enablers rather than chasing the highest-beta AI beneficiaries after a strong run. In other words, the trade is probably not “buy everything AI,” but “own the picks-and-shovels with pricing power and relative scarcity.”

Key Decisions for Investors

• Add to NVDA on any 3-5% pullback over the next 1-2 weeks; use it as the highest-conviction expression of AI capex durability, but size with a trailing stop given crowded positioning and valuation sensitivity.
• Overweight AVGO vs NVDA for the next 3-6 months via a long AVGO / short NVDA pair; thesis is that custom silicon, networking, and inference intensity diversify AVGO’s AI exposure and reduce single-name execution risk.
• Initiate a basket long in AI infrastructure suppliers (TSM, AVGO, AMAT, KLAC, MRVL) over the next month; risk/reward is attractive if the market keeps revising 2026 capacity plans higher, with downside limited to a demand pause rather than a structural break.
• Buy 6-12 month calls on TSM rather than common if seeking exposure; the payoff is asymmetry from sustained leading-edge pricing power, while premium outlay caps downside if capex spend temporarily compresses FCF.
• Avoid adding to high-multiple secondary AI beneficiaries until the next earnings cycle; the near-term trade is better in cash-generative enablers than in names whose bull case depends on multiple expansion rather than order visibility.