Investor brief · TSMC

The 2 nm ramp has entered the financial statements

Three conclusions separate real node progress from a headline that is easy to overread.

  1. 2 nm is commercial, but still earlyThe node supplied 3% of second-quarter wafer revenue. That is meaningful proof of shipment, not yet proof of mature economics.
  2. The ramp is visible in inventory before it is visible in marginInventory days rose from 80 to 87, which TSMC attributed primarily to the N2 ramp, while quarterly capital expenditure reached US$15.70 billion.
  3. The third-quarter margin range is the near-term testRevenue guidance rises sharply, yet the gross-margin range of 65% to 67% sits below the second quarter's 67.7%. Conversion matters more than demand alone.

TSMC's second-quarter report contains a number that changes the 2-nanometer discussion from roadmap analysis to financial analysis. N2 accounted for 3% of wafer revenue. The node is no longer only a promise attached to a production schedule; it is contributing to reported sales.

That milestone does not settle the investment case. A leading process can attract demand while its yield learning, tool depreciation, inventory build and overseas-fab dilution are still moving through the income statement. The useful question is therefore not whether customers want 2 nm. It is how quickly that demand becomes utilization, margin and free cash flow.

This article treats TSMC's July 16 disclosures as a reconciliation exercise rather than a price target. The company reported a very strong quarter. The tension lies inside the sequence of those numbers, not in denying their strength.

The headline is a transition, not a destination

Three readings define the handoff from demand to economics

Commercial proof

3%

of wafer revenue came from 2 nm

Scale already present

77%

of wafer revenue used 7 nm or more advanced nodes

Current profitability

67.7%

second-quarter gross margin

Second-quarter revenue reached US$40.20 billion, up 33.7% year over year and 12.0% sequentially. Net income rose 77.4% year over year. Those are not the figures of a company struggling to monetize demand. They are the figures of a company entering a new capacity cycle from an unusually strong base.

The base matters because it changes the burden of proof. When gross margin is already 67.7%, a successful 2 nm launch does not merely need to generate revenue. It eventually has to preserve an exceptional economic standard while TSMC absorbs the cost of new tools, new capacity and a broader manufacturing footprint.

TSMC 2 nm margin journey infographic showing early revenue mix, yield learning, volume absorption and the margin test
Read the path from top to bottom. Commercial demand appears first; mature unit economics arrive only after yield learning and volume absorb the fixed cost of the ramp.

The cost curve arrives before mature yields

The most revealing line in the management report is not a revenue line. Days of inventory increased from 80 in the first quarter to 87 days in the second, and TSMC said the increase was primarily due to the N2 ramp. Inventory is rising because the production system must be stocked, qualified and synchronized before the node reaches mature volume.

Capital expenditure makes the same transition visible in cash. TSMC spent US$15.70 billion in the quarter, up from US$11.10 billion in the first quarter. Operating cash flow increased, but capital spending increased faster, so free cash flow fell to NT$287.36 billion from NT$348.21 billion.

This distinction is crucial. A weaker free-cash-flow quarter caused by intentional capacity investment is not equivalent to deteriorating demand. But it still changes the valuation question: investors are temporarily funding a larger production base and need evidence that revenue, yield and pricing will catch up.

The mix is already doing two jobs

Second-quarter wafer revenue by process generation

The new node is a narrow leading edge attached to a much larger advanced-node profit engine.

2 nm · 3%

3 nm · 30%

5 nm · 33%

7 nm · 11%

Other · 23%

The 77% advanced-node share includes 2 nm, 3 nm, 5 nm and 7 nm. The visual proportions reproduce TSMC's reported mix rather than an estimate.

The mature 3 nm and 5 nm businesses provide the operating foundation that makes the N2 ramp investable. Together they represented 63% of wafer revenue. Their utilization and cost improvements can offset part of the early dilution from 2 nm, while 2 nm opens the next pricing and performance tier.

Platform mix points in the same direction. High-performance computing accounted for 66% of net revenue and grew 20% sequentially, while smartphone revenue declined 4% sequentially. The company is therefore not waiting for a handset recovery to finance the transition. HPC demand is carrying more of the production system.

That differs from the segment-dependence test used in our AMD Q2 analysis. TSMC is not asking whether one reporting segment can support the company. The question is whether a broad advanced-node base can fund a narrower and more expensive frontier without losing margin discipline.

The margin range is the signal

For the third quarter, TSMC guided to revenue of US$44.6 billion to US$45.8 billion. The midpoint implies another increase of roughly 12% from the second quarter. Gross margin, however, is guided to 65% to 67%, with a midpoint of 66%—below the 67.7% just reported.

That gap is not a contradiction. Revenue can accelerate while gross margin eases if the mix includes a larger share of early-ramp wafers, depreciation and start-up costs arrive ahead of mature output, foreign exchange moves, or overseas fabs dilute the consolidated result. TSMC explicitly said overseas fabs partially offset second-quarter cost improvement and higher utilization.

Signature module · Ramp cost reconciliation

Open the four gates between a wafer and durable margin

Each gate has its own evidence. A strong quarter can pass the demand gate while the cash gate is still absorbing the ramp.

Commercial adoption

Pass condition: 2 nm revenue continues to rise without requiring weakness in the 3 nm and 5 nm base. The first 3% establishes shipment, but the sequence of future mix disclosures will show the slope.

Manufacturing learning

Pass condition: utilization and cost improvement offset early-node waste. TSMC does not disclose a quarterly N2 yield figure, so investors should avoid inserting an invented number and use margin, mix and inventory as indirect evidence.

Inventory conversion

Pass condition: the seven-day increase associated with N2 stabilizes as shipments scale. Persistent inventory growth after the revenue mix expands would be a cleaner warning than the planned build itself.

Capital absorption

Pass condition: operating cash flow eventually grows faster than capital expenditure. One lower free-cash-flow quarter is compatible with a healthy ramp; a lasting gap would raise the return threshold.

The console matters because no single metric can prove the thesis. A rising 2 nm mix without inventory stabilization may describe shipment growth with poor conversion. A lower inventory ratio without mix growth may describe a slower ramp. Strong gross margin supported by mature nodes is welcome, but it does not by itself prove that N2 has reached attractive economics.

What would prove or break the thesis

Thesis strengthens

Revenue mix rises while the cash drag normalizes

Look for a larger 2 nm contribution, stable or improving inventory days, gross margin holding within the guided range and free cash flow recovering as the heaviest installation phase passes.

Thesis weakens

The ramp keeps consuming more than it converts

A larger N2 mix paired with persistent inventory expansion, repeated margin misses or a sustained capital-spending gap would imply that commercial progress is not yet producing the expected economic return.

There is also a concentration issue that deserves monitoring without turning it into a separate thesis. North America represented 78% of second-quarter revenue, and HPC represented 66%. The current ramp is being financed by powerful demand, but that demand is not evenly distributed across customers, regions or end markets.

The balanced conclusion is narrower than either the bullish or bearish shortcut. TSMC has already proved that 2 nm can ship into a record financial base. It has not yet had time to prove mature N2 economics. The next few quarters should be read as a conversion sequence: mix, inventory, margin and cash.

Questions about TSMC's 2 nm ramp

How much revenue did 2 nm contribute in the second quarter of 2026?

TSMC reported that 2 nm accounted for 3% of total wafer revenue. The figure establishes commercial contribution but is too early to represent a mature node mix.

Why did inventory days increase?

Inventory days rose from 80 to 87. TSMC said the increase was primarily due to the N2 ramp, which makes it a production-timing metric to follow rather than an automatic demand warning.

Why can revenue grow while gross margin eases?

Early-ramp cost, depreciation, product mix, currency and overseas-fab dilution can arrive before a new node reaches mature utilization. TSMC guides each line separately because the conversion is not instantaneous.

What is the most useful next checkpoint?

Read the next node-mix disclosure together with inventory days, gross margin and free cash flow. A higher 2 nm share is strongest when the other three measures show that the ramp is being absorbed.

Primary financial sources

TSMC results, management report and filings checked 18 August 2026

This analysis is informational and is not investment advice. Forward-looking guidance can change with demand, currency, utilization, execution and geopolitical conditions.