TSMC Q2 Revenue Hits a Record, But the Margin Test Comes Next. TSM Stock Valuation and Key Risks Explained

TSMC's Q2 revenue set a record, but the surprise was smaller than the headline suggests. Revenue of $39.62 billion (NT$1.270 trillion) was only 0.5% above the LSEG SmartEstimate. Meanwhile, a one percentage-point change in gross margin would affect quarterly gross profit by roughly NT$12.7 billion, almost three times the gross profit attached to the entire revenue beat. 

The investable question is no longer whether AI demand is strong. It is whether margins and guidance are strong enough to keep earnings estimates moving higher.

Let's break down what TSMC has confirmed, what the July 16 TSMC earnings report still needs to prove, and the three numbers that will determine whether TSM stock's current valuation is supported by earnings or stretched expectations.

TSMC’s Record Q2 Revenue Beat Estimates; But By Just 0.5%

TSMC reports revenue monthly, which means investors can calculate quarterly sales before the company publishes its complete income statement. The June release completed the April-to-June quarter:

Q2 revenue of around $39.62 billion ((NT$1.270 trillion) was inside TSMC's $39.0 billion to $40.2 billion guidance and slightly above the NT$1.264 trillion LSEG SmartEstimate compiled from 20 analysts. It increased 12% from Q1. First-half revenue reached NT$2.404 trillion, up 35.6% from the comparable 2025 period.

However, there is an exchange-rate wrinkle. TSMC gave its guidance in US dollars using an assumed rate of NT$31.7 per dollar. At that rate, the reported Taiwan-dollar revenue converts to about $40.08 billion, near the upper end of guidance. The final US-dollar figure will depend on TSMC's reported quarterly exchange rate.

The 67.9% June growth rate is eye-catching, but it should not be annualised. Monthly foundry sales can be lumpy because of shipment timing, customer product ramps and the comparison base. The quarterly total is the cleaner signal.

More importantly, this was a revenue release, not the Q2 earnings report. The following numbers remain unknown until July 16:

Analysts expect Q2 net profit to grow 58.8% year on year. That is a forecast, not a reported result. The gap between the known sales number and the unknown profit number is where the real investment debate sits.

Why TSMC’s Gross Margin Matters More Than Its Revenue Beat

TSMC's Q2 revenue exceeded the LSEG estimate by NT$6.381 billion. If that incremental revenue carried the 66.5% midpoint of management's gross-margin guidance, it would contribute about NT$4.24 billion of gross profit.

Now compare that with a one percentage-point change in gross margin across the whole quarter:

This is a gross-profit sensitivity, not an EPS forecast. Taxes, operating expenses and non-operating items are deliberately excluded. But it shows why Thursday's margin result matters more than Monday's sales headline. One margin point is worth roughly three times the gross profit associated with the entire revenue surprise.

Think of TSMC as a premium multiplex. Revenue is the value of every ticket sold. Gross profit is what remains after the direct cost of running those screenings. A packed theatre looks impressive, but an investor ultimately needs to know how much the operator earns per show. TSMC has already disclosed the packed house. July 16 reveals the economics behind it.

Q1 showed how powerful that economics can be. Revenue was $35.90 billion, slightly above guidance, but gross margin reached 66.2%, compared with management's 63% to 65% range. TSMC attributed the margin outperformance mainly to higher utilisation and better cost improvement.

The cash-flow numbers matter because TSMC is funding one of the semiconductor industry's largest capacity expansions while still generating substantial free cash flow. Management expects 2026 capital spending toward the upper end of its $52 billion to $56 billion range. In Q1 alone, it spent $11.1 billion. The demand mix also remained favourable. 

One important caveat: TSMC's HPC category is broader than AI accelerators. It also includes CPUs and other high-performance chips. Treating the entire 61% as pure AI revenue would overstate TSMC's disclosed AI exposure.

TSMC Is Earning More From Each Wafer, Not Just Shipping More

The strongest long-term evidence for TSMC is not a reported wafer price or a single month's growth. It is the economic value the company has extracted from each unit of production.

Using TSMC's disclosed wafer revenue and 12-inch-equivalent shipments, we can calculate a blended revenue-per-wafer proxy:

Between 2023 and 2025, wafer shipments increased 25.2%, while wafer revenue increased 73.8%. The calculated revenue-per-wafer proxy rose 38.9%. This proxy is not TSMC's disclosed average selling price. It combines changes in node mix, product complexity, customer mix and pricing. That is precisely why it is useful. It shows that TSMC's growth has come from selling economically more valuable manufacturing output, not merely pushing more wafers through its fabs.

Pricing power is part of that story, and it is supported by management commentary. In a June BBC interview, TSMC CFO Wendell Huang said inflation was increasing costs and did not rule out higher prices. He ruled out sudden fourfold or fivefold increases and said TSMC would reflect the value created by its technology leadership and manufacturing excellence. CEO C.C. Wei separately told shareholders he would like to raise prices, while rejecting abrupt increases similar to those seen in parts of the memory market.

That confirms the existence of pricing intent. The exact percentage remains unclear yet. TrendForce has cited industry reports suggesting approximately 3% to 10% increases for advanced nodes in 2026, with other reports placing the range at 5% to 10%. 

The same evidence standard applies to N2 wafer prices. TSMC has confirmed that N2 entered high-volume manufacturing in Q4 2025 with good yield, but it does not publish customer pricing. Reputed semiconductor publications have cited estimates around or above $30,000 per 300mm wafer, compared with roughly $20,000 for 3nm.

The sensible conclusion is that N2 commands a major premium, while its exact realised price remains confidential and likely varies by customer, design and volume. It is worth remembering that pricing does not flow directly to profit. Four counterweights matter:

• TSMC expects the initial N2 ramp to dilute full-year 2026 gross margin by 2 to 3 percentage points.
• Overseas fabs are expected to dilute gross margin by 2 to 3 points in their early stages, widening to 3 to 4 points later.
• Higher chemical, gas, electricity and equipment costs can absorb part of any price increase.
• Every 1% depreciation of the US dollar against the Taiwan dollar would reduce operating margin by approximately 0.3 percentage point, based on TSMC's 2025 results.

This creates the core tension: advanced nodes improve mix and pricing power, but their early ramps and global production footprint initially cost more. July's margins tell investors which force is currently winning.

TSMC's Valuation Requires Earnings to Compound Through 2028

The cleanest way to discuss whether expectations are reflected in the valuation is to examine the earnings path the market is being asked to accept. FactSet consensus estimates currently stand at NT$100.58 per Taiwan share for 2026, NT$129.97 for 2027 and NT$163.26 for 2028. 

One TSMC ADR represents five Taiwan shares. Using a recent ADR price of $434.11 and Reuters' NT$32.053 exchange rate gives this reverse-expectations framework:

This is not a price target. It holds the current price constant and asks what today's valuation would look like if consensus earnings materialise. Currency movements will change the ADR-equivalent figures.

The estimates imply a 27.4% EPS compound annual growth rate between 2026 and 2028. That is achievable only if AI and HPC demand remains strong, N2 ramps successfully, advanced packaging expands, and margin dilution stays manageable.

Expectations have already moved higher inside analyst models:

This is the most useful answer to the priced-in question. Analysts have already incorporated materially stronger earnings into their forecasts. The valuation can still receive support if July 16 triggers another round of estimate increases. If the report merely confirms numbers already in the models, the revenue record alone adds less new information than the headline suggests.

Wall Street Agrees on TSMC's Moat, but Not What It Is Worth

Analyst sentiment is overwhelmingly positive, but the assumptions underneath the ratings vary. FactSet's current compilation shows 38 Buy ratings, nine Overweight ratings and one Hold, with no Underweight or Sell ratings. That level of agreement is useful evidence of conviction, but it also means the optimistic thesis is hardly undiscovered.

Sources: Barclays, BofA and Susquehanna, TD Cowen, and TSMC Q1 analyst coverage.

The $400 to $625 range is not a menu of outcomes. It exposes the variables doing the work. The more bullish analysts assume that AI demand, scarcity and pricing will protect returns even as capex rises. The cautious view argues that excellent execution is already embedded in earnings expectations, leaving less room for operational disappointment.

Three TSMC Earnings Metrics That Could Change the Investment Thesis

Investors do not need to guess whether one record revenue headline settles the debate. They can use a three-part Revision Test while analysing TSMC’s Q2 earnings, to be reported later this week on July 16.:

1. Margin: Did gross margin finish above, within or below the 65.5% to 67.5% guided range?
2. Guidance: Is the midpoint of Q3 revenue guidance above or below the latest consensus available when the result is published?
3. Revision: Does management raise, maintain or reduce its full-year expectation of above 30% US-dollar revenue growth?

Five Key Risks to TSMC’s Bull Case

TSMC Stock Outlook After Record Revenue: What it Means for Investors

TSMC's Q2 revenue record confirms that demand for leading-edge manufacturing remains powerful. The longer-term evidence is stronger than one quarter: between 2023 and 2025, wafer revenue grew far faster than shipments, advanced nodes rose to 74% of wafer revenue, and gross margin expanded from 54.4% to 59.9%. Management has also confirmed that inflation is raising costs and that value-based pricing remains available, while industry reports point to premium pricing for N2.

The bull case is measurable: TSMC has strong utilisation, an expanding mix of advanced nodes, tight leading-edge and packaging capacity, and customers that continue to signal robust AI demand. If Q2 gross margin exceeds guidance and management raises its full-year outlook, consensus earnings could move higher again.

The bear case is equally concrete: The revenue beat was only 0.5%, consensus FY2026 and FY2027 EPS estimates have already risen 10.2% and 14.3% in three months, and the current valuation relies on approximately 27.4% EPS growth through 2028. N2, overseas fabs, currency and heavy capex all create ways for revenue growth to outpace profit growth.

The clean conclusion is not a buy, sell or hold call. It is an information rule: do not change the TSMC thesis because quarterly revenue set a record. Change it only if the July 16 margin, guidance and earnings revisions alter the company's 2026-28 profit path. Revenue has confirmed the demand. Profitability must now justify the expectations.