TSMC: Spending Its Lead To Keep It

TSMC earns an A- Gunderson Grade, ranking in the top 2% of tracked stocks and rated Buy. Its 2nm node contributes 8% of wafer revenue, supporting top-line growth but adding near-term margin pressure amid heavy investment; trailing 12-month EPS grew 41–63% year over year for four quarters, while gross margin declined 1.3 points due to the node ramp and overseas expansion costs.

Analysis

The key question is whether the new-node ramp converts revenue into durable incremental returns, not whether it adds sales. Early revenue can coexist with weak unit economics: yield learning, underutilization and overseas build-out may keep costs elevated before volume and pricing absorb them. The 1.3-point gross-margin decline is therefore a monitor, not proof of structural deterioration; equally, trailing EPS growth is backward-looking and does not establish that margins have bottomed.

Over the next 1–3 months, guidance on gross margin, capex and ramp execution should matter more than the grade or rating. Evidence of improving yields, utilization or pricing would support the view that near-term dilution is investment-led. Persistent margin erosion without stronger forward revenue conversion would challenge it. Over 6–18 months, successful 2nm scaling could reinforce TSMC’s differentiation and raise the cost of competing for leading-edge customers; Samsung and Intel Foundry face greater pressure to demonstrate credible alternatives. That benefit is conditional on execution, not guaranteed by node launch.

Contrarian angle: investors may treat the margin decline as a temporary ramp cost too readily. Revenue share alone does not reveal profitability, and continued investment can defer returns. Conversely, extrapolating one period of margin pressure into a lasting impairment could miss operating leverage as the node scales.

Key Decisions for Investors

• Keep TSM as a preferred semiconductor exposure, but avoid chasing on the rating alone. Consider adding on weakness only if subsequent guidance or results indicate gross-margin stabilization alongside continued advanced-node demand.
• For the next earnings cycle, track gross-margin guidance versus realized results, capex, and evidence of ramp efficiency. Treat rising revenue without margin improvement as a warning that scale economics are arriving later than expected.
• Use Samsung and Intel Foundry as competitive watchpoints rather than assuming a direct short: evidence of customer qualification or execution gains there would weaken TSMC’s pricing-power thesis; setbacks would support it.
• Falsification: reduce the constructive view if TSMC repeatedly guides or reports lower gross margins without a credible recovery path, or if ramp progress fails to translate into stronger revenue conversion over the next 1–3 quarters.