Bank of America expects TSMC to reach next year's capacity targets ahead of schedule, and it has raised its forecast for the 2nm process to 11% of revenue. That number is an analyst estimate, not company guidance. TSMC has not confirmed the timing shift itself. The claim rests on the bank's own channel checks and modeling.

The forecast carries real weight for the supply chain. TSMC leads advanced logic manufacturing. Its customers include Apple and Nvidia, both of which are expected to adopt the 2nm class node. Faster capacity installation would shorten their wait for wafers. It could also pull revenue forward into an earlier quarter than investors currently model. That last point is inference, not something BofA stated.

What does the 11% figure actually mean?

The number refers to revenue contribution, not wafer output. A process node can carry a high average selling price per wafer while running at modest volume. At the leading edge, that is the norm. New nodes price well above mature ones. So 11% of revenue may correspond to a smaller share of total wafer capacity. The bank's note, as reported by Moomoo, does not break this down. Readers should hold the two measures apart.

There is also a definitional caveat. "2nm process" in analyst shorthand typically covers TSMC's N2 family, including derivatives. The report does not specify whether the 11% covers the base node alone or the wider platform. That distinction changes the arithmetic materially. Base N2 alone at 11% would be a stronger signal than a blended figure.

Why does early capacity matter?

TSMC has told investors it would ramp 2nm in volume during 2025. Analysts have tracked that schedule through equipment deliveries and fab fit-out reports for months. A capacity target achieved ahead of schedule suggests demand from early customers is firm enough to justify pulling tool installations forward. Fabs do not install expensive lithography equipment on speculation. The equipment is ordered against committed demand.

But "ahead of schedule" is a soft phrase. The report does not say by how many weeks or quarters. A shift of one month has little commercial effect. A shift of one quarter would move meaningful revenue across a fiscal boundary. BofA's note, as summarized, does not quantify the gap. Without that number, the claim is directionally useful and precisely useless.

What is not said

Three things are missing from the report. First, no wafer capacity figure appears anywhere in it. Second, no yield data accompanies the forecast, and yield drives whether an early ramp translates into early revenue. Third, no pricing assumption is disclosed. BofA's 11% could assume aggressive wafer pricing, conservative volume, or both. Each combination tells a different story about 2026.

The competitive context matters here. Samsung Foundry and Intel both target the same advanced-node customers. If TSMC delivers 2nm capacity early and at good yields, rivals lose their window to win first-source positions at Apple and Nvidia. If yields disappoint, that window reopens. The report contains no yield evidence in either direction.

The bottom line

For now, the strongest confirmed fact remains modest: one major bank has raised its 2nm revenue estimate to 11% and expects early capacity achievement. Investors will learn whether the schedule holds when TSMC reports quarterly results and updates its own capital expenditure guidance, where actual numbers, not forecasts, will settle the question.