Tag: 2nm-technology

TSMC Q3 2026 Earnings: Growth Accelerates as 2nm Scales—and Cash Flow Strengthens

TSMC’s Q3 results show significant revenue growth, reaching US$46.72 billion, a 41.2% increase year-over-year. While operating income rose 75%, gross margins fell to 66.4% due to new technology costs. The 2nm segment now contributes 8% of wafer sales. Future valuations improved to approximately US$460 per ADR, reflecting strong demand.

TL;DR Summary

TSMC’s third-quarter results strengthened the business case behind its expansion. Leading-edge demand drove higher revenue and operating profit, 2nm gained a larger share of wafer sales, and free cash flow improved substantially.

The trade-off remains visible: new technology brings additional revenue but initially weighs on margins. Meanwhile, HPC accounts for an increasing proportion of the business.

  • Revenue reached US$46.72 billion, increasing 41.2% year over year.
  • Operating income increased 75%, while gross margin declined sequentially to 66.4%.
  • 2nm represented 8% of wafer revenue, up from 3% in Q2.
  • Free cash flow increased to NT$476.56 billion, from NT$287.36 billion.
  • Our updated probability-weighted forward value is approximately US$460 per ADR, versus US$440 previously.

The business outlook has improved. Whether the shares offer an attractive entry point still depends on the price paid and the investor’s tolerance for capital intensity and geopolitical risk. investor.tsmc.com

Quarter Recap

TSMC reported Q3 revenue of NT$1,494.24 billion, equivalent to US$46.72 billion. Revenue increased 50.9% year over year in NT dollars and 41.2% in US dollars.

Net income attributable to shareholders reached NT$763.06 billion, increasing 68.7%. Diluted earnings were NT$29.43 per ordinary share, or US$4.60 per ADR.

Revenue exceeded management’s previous US$44.6–45.8 billion guidance. Gross margin of 66.4% was within guidance, while operating margin of 58.7% exceeded the projected 56–58% range. investor.tsmc.com

Sequential comparisons require care. Q2 included Vanguard-related investment gains that did not recur in Q3. Consequently, reported net income increased 8% sequentially, while operating income increased 14.3%. The two measures describe different aspects of the quarter rather than contradictory performance. investor.tsmc.com

Key Highlights

Leading-edge demand continues to expand

The 2nm contribution increased from 3% to 8% of wafer revenue, while 3nm increased from 30% to 33%. Technologies at 7nm and below collectively represented 79%.

HPC revenue increased 22% sequentially and accounted for 68% of total revenue. Smartphone and automotive revenue also increased sequentially, by 10% and 23%, respectively. HPC remains a broader category than AI alone; its entire revenue should not be described as AI revenue. investor.tsmc.com

These figures support our interpretation that TSMC’s growth is being realized through shipments and technology adoption—not simply through expectations about future demand.

Cash conversion improved

Operating cash flow rose to NT$919.77 billion, while capital expenditure declined sequentially to NT$443.21 billion. Free cash flow consequently increased to NT$476.56 billion.

Cash and marketable securities reached NT$4,016.98 billion. investor.tsmc.com

This updates an important part of our Q2 assessment. Capital intensity remains a concern, but the latest quarter does not support describing cash conversion as continuously deteriorating.

Growth continues alongside ramp costs

Management attributed the sequential gross-margin decline primarily to the initial N2 ramp.

For Q4, TSMC expects revenue of US$52.3–54.3 billion, gross margin of 65–67%, and operating margin of 56.5–58.5%. Its presentation forecasts approximately 44% full-year US-dollar revenue growth. investor.tsmc.com

Our interpretation is that demand remains strong, but investors should not assume that revenue growth will translate into proportional margin expansion.

Market Reaction

Editorial note: Exchange-only verification of the October 15 price session remains outstanding. This draft therefore makes no claim about the closing reaction or investors’ reasons for trading. This section should be completed before publication.

SWOT Analysis

TSMC combines technology leadership and substantial cash generation with a business model that requires large, continuing investment.

The percentage ranges below are illustrative changes in modeled equity value, not predicted stock-price movements. Each is sensitivity-derived from explicit assumptions. They overlap and must not be added together.

Strengths

Leading-edge technology mix: +3% to +5% sensitivity. More advanced production can support revenue per wafer and customer retention. Assuming technology mix produces 3–5% more revenue than our base forecast, with unchanged net margin and valuation multiple, modeled value increases by the same proportion. The financial benefit depends on pricing covering the higher manufacturing costs.

Financial flexibility: +2% to +4% sensitivity. Cash generation provides capacity to fund expansion and absorb fluctuations without issuing equity. If that flexibility supports a 23.5–24× earnings multiple rather than our 23× reference, modeled value increases approximately 2–4%. This premium is a judgment, not an automatic consequence of holding cash.

Weaknesses

Ramp-related margin pressure: −2% to −4% sensitivity. New-node production can increase revenue while temporarily reducing profitability. Against an illustrative 50% normalized net margin, a one- to two-percentage-point reduction lowers earnings and value by 2–4%, assuming unchanged revenue and multiple.

HPC concentration: −4% to −9% sensitivity. An increasingly HPC-dependent revenue mix raises exposure to customer investment cycles. Applying a one- to two-turn discount to the 23× reference multiple reduces value approximately 4–9%, even if earnings meet expectations.

Opportunities

Further demand expansion: +5% to +10% sensitivity. Continued leading-edge demand could lift shipments above our forecast. Revenue 5–10% above the base case produces an equivalent increase in value if margins and the multiple remain unchanged. This overlaps with the technology-mix opportunity.

Successful 2nm maturation: +2% to +4% sensitivity. Yield improvements and manufacturing efficiency could improve earnings conversion after the initial ramp. A one- to two-percentage-point improvement above the illustrative 50% net margin raises modeled earnings by 2–4%. This is an alternative to the margin-pressure outcome, not an additional benefit to count alongside it.

Threats

Demand falling short of capacity: −13% to −24% sensitivity. Lower utilization could coincide with rising depreciation. A stress combining revenue 5–10% below forecast, net margins of 48–46%, and multiples of 22–21× produces this range relative to the 50% margin and 23× reference.

Multiple compression: −9% to −17% sensitivity. Even satisfactory business performance may not protect the shares if investors pay less for future earnings. A decline from 23× to 21–19× produces this valuation effect with earnings unchanged.

Taiwan-related disruption: losses are not reliably bounded. A three- to five-turn multiple discount would reduce modeled value approximately 13–22% without changing earnings. However, a blockade, conflict or prolonged production interruption could impair earnings as well. That more severe tail risk cannot responsibly be contained within an ordinary multiple sensitivity.

Valuation Scenarios

Our valuation uses 2027 earnings to estimate a 12-month forward value. It is not a discounted-cash-flow calculation of present intrinsic value.

The model uses only company disclosures as financial inputs. Future growth, margins, valuation multiples and probabilities are our independent assumptions—not management guidance or analyst consensus.

We estimate adjusted 2026 earnings of approximately US$17.14 per ADR, compared with US$16.40 previously. This combines reported nine-month earnings with a Q4 model based on guidance midpoints and removes the previously identified Vanguard gain. It does not remove every potentially unusual item.

Bear Case — Approximately US$305

Revenue grows 5% in 2027, but net margin declines to 45%. Modeled ADR earnings fall to approximately US$16.03.

A 19× multiple produces a value near US$305. We assign a 25% probability.

This scenario reflects slower demand and weaker manufacturing economics. It does not represent a severe geopolitical disruption.

Base Case — Approximately US$465

Revenue grows 22%, while net margin settles at 49%. Modeled ADR earnings reach approximately US$20.27.

A 23× multiple produces a value near US$465. We assign a 50% probability.

This assumes continued leading-edge expansion, but allows for manufacturing costs to constrain profitability.

Bull Case — Approximately US$595

Revenue grows 30%, with net margin reaching 52%. Modeled ADR earnings rise to approximately US$22.93.

A 26× multiple produces a value near US$595. We assign a 25% probability.

This requires sustained demand, strong utilization and effective execution together.

Using unrounded scenario values, the probability-weighted result is approximately US$458, rounded to US$460 per ADR. Scenario headlines are rounded to the nearest US$5.

Verdict

TSMC’s Q3 report strengthens our growth thesis. Advanced-node adoption increased, earnings expanded, and cash conversion improved. The initial 2nm margin cost is visible, but it has not prevented substantial profit growth.

We therefore raise our forward valuation from US$440 to approximately US$460, without increasing the scenario multiples. The upgrade comes from earnings expectations rather than a willingness to pay more for each dollar of profit.

That distinction matters. TSMC can remain an exceptional business while its shares offer an insufficient valuation cushion.

A price of US$368 would represent a 20% discount to the rounded forward estimate. That is an arithmetic threshold—not a guaranteed safe entry point—and a present-value assessment would additionally account for the required investment return.

For growth investors, the central question remains whether demand will persist long enough to generate attractive returns on the capacity being built. Q3 provides stronger evidence of current demand and financing capacity, but it cannot settle that longer-term question.

Call to Action

Which matters more to your TSMC thesis: the pace of leading-edge growth, or the returns earned on the investment needed to deliver it?

Share your perspective in the comments.

Disclaimer

This article is for informational and educational purposes only and is not investment advice or a recommendation to buy or sell securities. Valuations and sensitivities depend on assumptions that may not materialize. Semiconductor cyclicality, manufacturing execution, currencies and geopolitical events could cause materially different outcomes.

TSMC Q2 2026 Earnings: AI Demand Accelerates, but Is the Stock Still Attractive at $410?

TSMC reported strong Q2 results with revenues of $40.20 billion, driven by high demand for AI and advanced chips. Despite exceeding forecasts, shares fell 2.3% due to high capital spending and margin pressures. The company raised its 2026 revenue growth outlook, but geopolitical risks remain a concern for investors.

TL;DR Summary

TSMC delivered another exceptional quarter, driven by leading-edge chips and rapidly expanding AI demand. Revenue reached the top of management’s guidance, margins exceeded expectations, and the company raised its full-year growth outlook. Yet the ADR fell 2.3% after the report.

The problem was not weak execution. It was the combination of a demanding valuation, a much larger capital program, and visible margin pressure from the 2nm ramp and overseas fabs.

  • Q2 revenue increased 33.7% year over year in US dollars to $40.20 billion.
  • Gross margin reached 67.7%, while operating margin reached 60.3%.
  • High-performance computing grew 20% sequentially and represented 66% of revenue.
  • Management now expects 2026 revenue growth to be slightly above 40% in US-dollar terms.
  • The 2026 capital budget was raised to $60–64 billion, increasing both TSMC’s growth capacity and its execution risk.
  • Our estimated fair value is $440 per ADR, compared with the July 16 closing price of $409.74.
  • The shares appear slightly undervalued, but the margin of safety remains limited.

Quarter Recap

TSMC reported Q2 2026 revenue of NT$1.270 trillion, equivalent to $40.20 billion. Revenue increased 36.0% year over year in New Taiwan dollars and 33.7% in US dollars. Sequential growth was 12.0%.

Profitability was even stronger. Gross profit increased 57.2% year over year, operating income increased 65.4%, and net income attributable to shareholders increased 77.4%. Gross margin reached 67.7%, operating margin reached 60.3%, and annualized return on equity reached 45.9%.

The headline EPS figure requires one adjustment. Reported ADR earnings of $4.31 included approximately $0.35 per ADR from gains related to Vanguard International Semiconductor shares. Excluding that identified contribution, normalized Q2 ADR earnings were approximately $3.96.

The underlying operating result was still exceptional, but the full 77.4% increase in reported EPS was not entirely recurring.

For Q3, management expects revenue of $44.6–45.8 billion. The midpoint represents approximately 12% sequential growth and 37% year-over-year growth. Gross margin is expected to fall to 65–67%, while operating margin is expected to moderate to 56–58%.

The market reacted negatively despite the strong report. TSM closed at $419.48 before the release and $409.74 afterward, a decline of 2.32%. The ADR opened down more than 3% before recovering part of the loss.

This was not a rejection of TSMC’s operating performance. It was a reassessment of what investors must pay—and how much TSMC must spend—to capture the AI opportunity.

Key Highlights

  • AI has become the central growth engine. HPC revenue increased 20% sequentially and represented 66% of total revenue, up from 61% in Q1 and 60% a year earlier.
  • Advanced-node adoption continues to strengthen. Technologies at 7nm and below generated 77% of wafer revenue. The mix included 3% from 2nm, 30% from 3nm, 33% from 5nm, and 11% from 7nm.
  • The 2nm ramp has begun. Its first revenue contribution confirms that the technology is entering commercial production, but management expects the ramp to reduce second-half gross margin by approximately three to four percentage points.
  • Capital expenditure is rising sharply. TSMC raised its 2026 capital budget to $60–64 billion from an original $52–56 billion. Management attributed the increase mainly to stronger customer demand and, secondarily, equipment-price inflation.
  • Free cash flow remains positive but is under pressure. Operating cash flow reached NT$783.36 billion, while capital expenditure reached NT$496.00 billion. Free cash flow was NT$287.36 billion, down sequentially because investment grew faster than operating cash generation.
  • The balance sheet remains formidable. TSMC ended the quarter with NT$3.518 trillion in cash and marketable securities and NT$2.486 trillion in net cash reserves.
  • Demand outside AI is less convincing. Management described consumer and price-sensitive markets as challenged. Mature-node shortages remain concentrated in AI-related power-management chips and sensors rather than the broader market.
  • Advanced packaging remains a bottleneck. Management acknowledged that packaging capacity is still tight enough to restrict customer growth, even as TSMC increases back-end investment.
  • The long-term technology road map remains intact. A14 risk production is scheduled for 2027 and volume production for 2028. Compared with N2, A14 is designed to deliver 10–15% more speed at the same power or 25–30% lower power at the same speed.

SWOT Analysis

TSMC combines one of the strongest competitive positions in global technology with unusually concentrated geopolitical and capital-allocation risks. The estimated price-impact ranges below represent isolated 12-month sensitivities and should not be added together because several factors overlap.

Strengths

  • Leading-edge manufacturing leadership: +8% to +14%. Advanced technologies represented 77% of wafer revenue, while 2nm entered commercial production. This mix supports pricing, customer retention, and long-term returns on capital.
  • Exceptional AI and HPC demand: +10% to +18%. HPC now generates two-thirds of revenue, and management continues to receive strong demand signals from customers and cloud-service providers.
  • Industry-leading profitability: +5% to +9%. A 67.7% gross margin and 60.3% operating margin demonstrate considerable pricing power, utilization, and manufacturing efficiency.
  • Strong balance sheet: +3% to +7%. TSMC’s net-cash position allows it to finance an enormous expansion program without immediately compromising financial stability.
  • Deep customer integration: +4% to +8%. Management estimates that technology selection, qualification, and capacity preparation can take approximately five years. This makes foundry switching slow, costly, and operationally risky for customers.

Weaknesses

  • Rising capital intensity: −6% to −12%. The $60–64 billion capital budget reduces near-term free-cash-flow conversion and raises the financial cost of any future forecasting error.
  • Margin dilution: −5% to −10%. The 2nm ramp will pressure second-half margins, while overseas fabs are expected to reduce gross margin by two to three percentage points initially and three to four points at a later stage.
  • HPC and geographic concentration: −4% to −8%. HPC accounts for 66% of revenue, while North American customers account for 78%. This increases exposure to a concentrated AI investment cycle.
  • Advanced-packaging constraints: −3% to −7%. Packaging shortages currently limit how quickly customers can translate leading-edge wafers into complete AI systems.
  • Nonrecurring support to reported EPS: −1% to −3%. The Vanguard-related gain added NT$2.24 per ordinary share to reported earnings and should not be capitalized as recurring income.

Opportunities

  • The 2nm ramp: +8% to +15%. Initial margin dilution could give way to stronger earnings once utilization, yield, and cost improvements develop.
  • Agentic AI and CPU demand: +6% to +12%. Agentic systems require CPUs alongside accelerators. Major x86, Arm, and RISC-V participants are predominantly TSMC customers, allowing the company to benefit across competing architectures.
  • Advanced-packaging expansion: +5% to +10%. Closing the supply-demand gap in CoWoS and related technologies would allow more existing AI demand to convert into shipments.
  • The A14 technology family: +4% to +8%. A14, A13, and A12 could extend TSMC’s leading-edge advantage through the end of the decade.
  • Global capacity: +3% to +7%. Manufacturing expansion in the US, Japan, and Europe may deepen strategic customer relationships and reduce concerns about geographic supply concentration.

Threats

  • An AI infrastructure slowdown: −15% to −25%. The investment program assumes that customer demand remains strong for several years. A slowdown could expose TSMC to lower utilization just as depreciation rises.
  • Taiwan geopolitical risk: −20% to −35%. This remains the largest low-frequency, high-impact risk to the investment case and warrants a persistent valuation discount.
  • Simultaneous expansion risk: −8% to −15%. TSMC must ramp 2nm, expand packaging, and execute several overseas projects without losing its manufacturing discipline.
  • Weakness outside AI: −5% to −10%. Consumer, price-sensitive, and most commodity mature-node categories are not participating equally in the current expansion.
  • Foundry and packaging competition: −6% to −12%. Rapid displacement is unlikely, but selective customer diversification or competing packaging solutions could reduce TSMC’s share of future value creation.

Valuation Scenarios

Our valuation starts with estimated normalized 2026 ADR earnings of approximately $16.40. This removes the identified Vanguard gain and models the second half from TSMC’s Q3 guidance and full-year growth outlook.

We then estimate 2027 earnings under three operating environments. These are independent scenarios, not analyst-consensus forecasts.

  • Bear case — $327, with a 25% probability. Normalized 2027 ADR earnings rise only 5% to approximately $17.20. Applying a 19× multiple produces $326.80, rounded to $327. This scenario reflects weaker AI spending, lower utilization, and continued overseas-fab dilution. It implies approximately 20% downside from $409.74.
  • Base case — $445, with a 50% probability. Normalized 2027 ADR earnings increase 18% to approximately $19.35. A 23× multiple produces $445.05, rounded to $445. This assumes that 2nm and packaging capacity expand successfully while overseas costs prevent meaningful margin expansion. It implies approximately 9% upside.
  • Bull case — $546, with a 25% probability. Normalized 2027 ADR earnings increase 28% to approximately $20.99. A 26× multiple produces approximately $546. This requires continued accelerator demand, incremental CPU demand from agentic AI, strong utilization, and effective cost control. It implies approximately 33% upside.

Weighting the three scenarios produces a value of approximately $441 per ADR. Our final fair-value estimate is rounded to $440, with a reasonable range of $400–480.

The estimated 2026 dividend of NT$24 per ordinary share is equivalent to roughly $3.75 per ADR at an exchange rate near NT$32. Including the dividend, a move from $409.74 to fair value would produce a potential total return of approximately 8%.

Verdict

TSMC’s Q2 report strengthened the business thesis. AI demand is translating into revenue, advanced-node utilization, and a larger capital program. The company is not investing against a vague market narrative; management is responding to direct customer demand and acknowledged capacity shortages.

That does not automatically make the ADR inexpensive. At approximately $410, investors are paying for substantial growth while accepting three important risks: unusually high capital expenditure, margin dilution, and geopolitical concentration.

Our $440 fair value suggests that the ADR is modestly undervalued, but the expected return is not large enough to constitute a clear margin-of-safety opportunity. The stock may still suit a tech-savvy growth investor who is confident in the durability of AI infrastructure spending and willing to tolerate geopolitical volatility. It is less compelling for an investor seeking a classic value entry point.

The central question is no longer whether TSMC is benefiting from AI. The evidence clearly says it is. The question is whether AI demand will remain strong enough, for long enough, to generate attractive returns on the enormous capacity now being built.

Call to Action

Would you buy TSMC near $410, or would the combination of capital intensity and geopolitical risk require a larger discount?

Share your view in the comments, and follow SWOTstock for independent, evidence-based analysis of the next major earnings report.

Disclaimer

This article is for informational and educational purposes only and does not constitute financial advice, investment advice, or a recommendation to buy or sell any security. Valuation estimates depend on assumptions that may not materialize.

Semiconductor demand, foreign-exchange movements, capital expenditure, technology execution, and geopolitical events may cause actual outcomes to differ materially. Investors should conduct their own research and consider their objectives, financial circumstances, and risk tolerance before making an investment decision.