BofA raises TSMC stock price target to $500 on strong AI demand

BofA Securities raised its price target on TSMC to NT$2,530 from NT$2,360 and reiterated a Buy rating, citing 7% to 9% Q/Q sales growth expected for Q2 2026 and gross margin expansion to about 66% from 63% to 65% guided in Q1. The firm highlighted strong HPC and AI demand, 3nm utilization above 100%, and premium pricing as key margin drivers despite 2nm ramp and summer electricity cost headwinds. Recent revenue data also showed March 2026 sales up 45.2% year over year, reinforcing a constructive fundamental outlook.

Analysis

The near-term market is still underestimating how much of TSMC’s value is now coming from bottleneck pricing power rather than simple unit growth. When leading-edge nodes run at or above full utilization, incremental demand migrates from volume sensitivity to allocation sensitivity, which tends to preserve margins even if broader semiconductor cyclicality softens. That makes TSM not just an AI beneficiary, but a toll booth on scarce advanced capacity; the second-order winner is its ecosystem of equipment, packaging, and specialty materials suppliers that get pulled forward by every incremental capacity decision.

The bigger strategic implication is that this strength may persist even if AI spending normalizes, because customers are optimizing for wafer-turn time and supply assurance, not just die cost. That reduces the odds of a classic semiconductor downcycle in TSM first and pushes the adjustment burden onto less differentiated foundry peers and memory names with weaker pricing power. The risk is that investors extrapolate current utilization and ignore the margin drag from the next node transition: if the 2nm ramp becomes capital- or yield-intensive, the market could see a few quarters of flat earnings power despite healthy top-line growth.

Consensus is likely missing how much of the re-rating has already happened versus how much is still being justified by forward earnings compounding. At this point, the stock is less about cheapness and more about whether 2027 estimates prove conservative; if AI order visibility stays intact, estimate revisions can carry the name higher without multiple expansion. But if summer power costs, geopolitical friction, or customer concentration create even a modest miss, the stock can de-rate quickly because expectations are now high and ownership is crowded. The cleanest contrarian read is that downside is not from demand collapse, but from a transition-year pause that is easy to overlook in an otherwise strong tape.

Key Decisions for Investors

• Maintain/accumulate long TSM on 1-3 month pullbacks, using weakness toward the 50-day as entry; target is continued estimate revisions over the next 2 quarters, with downside primarily from 2nm ramp execution rather than demand.
• Pair long TSM vs short a basket of weaker foundry/legacy semiconductor names over the next 6-12 months to express divergence in pricing power and utilization; thesis is that scarce leading-edge capacity stays tight while lower-quality peers absorb any cyclical slowdown.
• Buy TSM upside calls or call spreads into the next earnings/guide event for 3-6 month exposure; risk/reward favors convexity because forward revisions, not multiple expansion, are the main upside driver.
• If already long, trim only into sharp extension rather than on the headline strength; the crowding risk is real, but the better exit trigger is any sign that 2nm costs or customer mix flatten gross margin progression.