TSMC April sales jump 17.5% on sustained AI chip demand

TSMC reported April revenue of NT$410.73 billion, up 17.5% year over year, driven by continued demand for AI chips, though revenue fell 1.1% from March. For the first four months of 2026, revenue rose 29.9% to NT$1.54 trillion, and the company reiterated current-quarter sales guidance of $39 billion to $40.2 billion. Shares in Taipei were down 0.9% as investors weighed strong AI demand against U.S.-China tech restrictions and broader supply-chain/geopolitical risks.

Analysis

TSMC is still the cleanest public-market expression of the AI capex cycle, but the real signal is that demand is broad enough to absorb both a strong absolute print and sequential noise. That matters because the next leg of upside is less about unit growth and more about pricing power in leading-edge nodes, advanced packaging, and tight capacity allocation; if those remain constrained, gross margin mix can stay elevated even if top-line growth normalizes. For NVDA, this reinforces the durability of the compute buildout, while AAPL benefits only indirectly through supply assurance rather than any near-term demand reacceleration.

The second-order risk is that a strong revenue cadence invites the market to over-extend expectations into the June/September quarters just as geopolitical and export-control headlines can force customers to diversify supply chains. The most vulnerable piece is not AI demand itself, but the concentration of that demand in a single foundry ecosystem; any incremental friction on cross-strait logistics, licensing, or tool delivery would hit sentiment first and volumes later. In that sense, the stock reaction can stay choppy even if fundamentals remain intact.

The contrarian read is that investors may be underestimating how much of the AI supply chain has already de-risked around TSMC, making the near-term setup less about a demand surprise and more about execution against a high bar. If management guides within or slightly above the current range, the upside to the equity may be capped unless there is evidence of margin expansion or a material capex inflection from hyperscalers. That creates a favorable setup for relative-value expressions rather than outright beta longs.

Key Decisions for Investors

• Long TSM vs. short SOXX for 4-8 weeks: own the highest-quality foundry bottleneck while fading broader semi beta that is more exposed to guidance dispersion; target a 5-8% relative outperformance if AI spend remains concentrated.
• Buy NVDA on dips only, with a 1-2 month horizon: treat this as confirmation of supply durability rather than a reason to chase; upside is strongest if TSM’s next update signals tighter advanced-node capacity, while downside is a sentiment reset if sequential growth slows.
• Initiate a short-dated call spread in TSM if the stock re-rates sharply on the print: the asymmetry is toward multiple compression after a good-but-not-revisionary update, with theta working in your favor over the next 30-45 days.
• Use AAPL as a hedge inside a semiconductor/AI basket: small positive read-through from stable supply is not enough for a standalone long, but it can offset idiosyncratic risk if supply-chain headlines worsen over the next quarter.
• Watch for any commentary on capex and advanced packaging capacity over the next earnings cycle; if spend inflects upward, rotate from broad semi exposure into TSM/NVDA pair trades because the bottleneck theme would extend another 2-3 quarters.