Buy TSMC stock as ’recent concerns are overdone’: BofA

Bank of America reiterated a Buy rating on TSMC with a NT$2,560 price target, arguing competitive concerns around foundry share are overdone. The note highlights TSMC’s 25% CAGR capacity growth in 3nm/5nm from 2022 to 2027, 70% CAGR N2 capacity growth from 2026 to 2028, and CoWoS/SoIC expansion at 80%/90% CAGR through 2027. BofA also said TSMC’s CoWoS yield is above 98%, versus Intel EMIB-T at 80%-85% pilot yield, reinforcing the company’s advanced-node and packaging lead.

Analysis

The market is still underestimating how much TSMC’s scaling advantage compounds into customer lock-in. The key second-order effect is not just share retention at the leading edge, but the widening gap in ecosystem confidence: once OEMs and hyperscalers anchor design roadmaps to the most reliable capacity, switching costs rise sharply because packaging, yield learning, and software validation all become path-dependent. That dynamic should keep premium allocation flowing to TSMC even if headline wafer pricing softens, and it also supports a higher-quality mix shift that protects margins through the next cycle.

The negative read-through is disproportionate for Intel. Advanced packaging is becoming the bottleneck, and if Intel cannot close its yield gap quickly, its foundry narrative risks a classic “credibility trap” where customers test samples but delay meaningful volume until the execution proof is undeniable. That creates a window over the next 12-24 months where capex intensity rises faster than revenue conversion, pressuring FCF and increasing the probability of further strategic concessions or external partnerships.

Apple is partially insulated in the near term, but the more important implication is strategic optionality: TSMC’s capacity growth reduces the odds that Apple is forced into a supplier diversification decision for performance reasons. The contrarian takeaway is that the market may be too focused on whether rivals can match node specs, when the real moat is time-to-yield and packaging throughput. If TSMC sustains these execution deltas, the earnings surprise is likely to come from mix and utilization, not just unit growth.

Key Decisions for Investors

• Add to TSM on weakness over the next 1-2 sessions; favor a 3-6 month horizon. Risk/reward remains attractive because execution premium can re-rate the stock further if the market starts pricing in sustained node and packaging leadership.
• Short INTC as a relative-value hedge against TSM over the next 3-9 months. The trade works best if Intel’s foundry narrative continues to require proof points while capex stays elevated; stop if mass-production yields improve materially ahead of schedule.
• Pair trade: long TSM / short INTC. This captures the widening credibility gap in yield and advanced packaging execution without relying on broad semiconductor beta.
• Buy medium-dated TSM calls or call spreads into any post-event consolidation. Best setup is 2-4 months out, targeting upside from renewed confidence in capacity ramp and margin durability; keep upside-defined due to Taiwan/geopolitical headline risk.
• Avoid initiating a fresh directional AAPL short on this headline; instead treat any Apple weakness as a hedge opportunity for short-term semiconductor sentiment trades. The supplier concentration risk is reduced, not increased, by TSMC’s current execution trajectory.