Morgan Stanley Raises TSMC Target Price! Can AI Expansion Sustain Margins?

The artificial intelligence (AI) boom is redefining TSMC's growth trajectory. Following TSMC's (2330-TW) Q2 earnings release, Morgan Stanley raised its target price from NT$2,888 to NT$2,988 and upgraded its three-year profit forecast, citing sustained demand for advanced process technologies and advanced packaging. However, higher capital expenditures and margin pressures from 2nm mass production remain critical market watchpoints.

After TSMC announced its Q2 financial results and raised its full-year outlook on Thursday (16th), Morgan Stanley promptly increased its target price to NT$2,988 from NT$2,888, maintaining an 'Overweight' rating.

Analysts noted that AI demand strength has exceeded prior market expectations, reshaping TSMC's 2026 revenue growth pace and capital expenditure rhythm. Yet, analysts caution this is not an unreservedly optimistic report.

TSMC's Q2 financials revealed quarterly revenue of NT$1.27038 trillion, a 36.0% year-on-year increase; gross margin reached 67.7%, operating margin 60.3%, and earnings per share (EPS) NT$27.25.

For Q3, the company forecasts revenue between USD 44.6 billion and USD 45.8 billion, with gross margin between 65% and 67%, and operating margin between 56% and 58%.

More notably, TSMC revised its full-year outlook. The company raised its 2026 USD-denominated revenue growth forecast from 'over 30%' in April to 'slightly above 40%.' Management described AI demand as 'extremely strong' during the earnings call, driven by robust order signals from cloud service providers and their downstream customers.

It's important to note that this growth momentum is not due to a broad semiconductor recovery. TSMC management acknowledged that non-AI demand remains weak, with inventory corrections ongoing in consumer electronics and traditional semiconductors.

AI-related orders are sustaining advanced process capacity utilization because this segment contributes disproportionately to TSMC's overall revenue and profitability compared to other applications.

Morgan Stanley also raised its long-term AI semiconductor revenue CAGR assumption from TSMC's previously mentioned 'mid-50s%' to 70%–80%, although TSMC did not formally announce a new long-term growth rate during the call.

Capital Expenditure to Reach $62 Billion, Advanced Processes Remain Core

The most eye-catching figure in this revision is the significant increase in capital expenditure guidance.

TSMC raised its 2026 capex target to $60–64 billion, with a midpoint of approximately $62 billion, of which 70%–80% will be allocated to advanced processes.

Morgan Stanley also revised upward its capex assumptions for TSMC over the coming years, expecting continued expansion of 2nm and 1.6nm capacity.

In addition to 2nm, TSMC mentioned during the call that the scale of A14 and its derivative technologies will exceed N2, with longer product lifecycles. U.S. fab capacity planning is also part of its long-term investment strategy.

However, Morgan Stanley cautioned that higher capex does not equate directly to proportional increases in usable capacity, as part reflects rising equipment prices and prepayment impacts. This indicates that rising investment intensity reflects not only urgent demand but also changes in cost structures and payment timing.

Morgan Stanley's long-term assumption remains optimistic, expecting capital intensity to decline to around 30% by 2028 as AI-related revenue scales.

Target Price Implies 20x P/E, with Comprehensive Earnings Upgrades

Morgan Stanley raised TSMC's target price from NT$2,888 to NT$2,988, primarily due to higher multi-year earnings estimates rather than a one-off earnings surprise.

The firm revised upward TSMC's 2026–2028 revenue, gross profit, net profit, and EPS forecasts. The new target price corresponds to approximately 20x P/E for 2027; at current share prices, it's about 17x, close to the historical average of ~16.5x since 2018.

Morgan Stanley identifies two key factors supporting higher valuation:

First, the potential for continued upward revisions in AI-related revenue;

Second, TSMC's pricing power and capacity utilization in advanced processes.

The report expects 5%–10% room for wafer price increases in advanced nodes by 2027, as customer dependency on cutting-edge nodes will sustain TSMC's pricing power.

The market's willingness to assign TSMC a higher valuation stems from long-term AI-driven demand. With orders continuously increasing, advanced process nodes, CoWoS and other advanced packaging, and future node capacities remain fully utilized.

With strong demand for high-end AI chips, supply tightness is expected to support sustained growth in TSMC's revenue and profitability.

Gross Margin Realization Speed Remains Key Variable

Despite the optimistic outlook, whether gross margins can be sustained remains the core market concern.

TSMC's Q3 gross margin guidance midpoint is at a historical high, but below some market expectations nearing 70%.

As 2nm process enters early mass production, depreciation, yield ramp-up, and upfront investments will create cost pressures. Overseas production costs are generally higher than in Taiwan, so faster advanced-node expansion could further suppress short-term margins.

Morgan Stanley remains relatively optimistic on medium-to-long-term margins, believing that despite significantly higher depreciation from advanced-node expansion over the next three years, revenue scale growth, product mix optimization, and pricing power at advanced nodes could partially offset cost pressures. In the long run, 60% could serve as a support level for TSMC's gross margin.

Additionally, weak non-AI demand poses another constraint. Morgan Stanley noted that China smartphone shipment tracking data is weaker than expected, suggesting TSMC's outlook upgrade should be interpreted as AI-driven, not a signal of broad semiconductor recovery.

On competition, if Samsung and Intel (INTC-US) Foundry narrow the process node gap in future nodes, it could impact long-term assessments of TSMC's market share, pricing power, and profitability.

Morgan Stanley concluded that AI demand strength is sufficient to drive TSMC's simultaneous upgrade of 2026 growth expectations and capex plans. However, whether the higher target price gains sustained market acceptance ultimately depends on whether these orders translate into high capacity utilization at 2nm and beyond, stronger pricing power, and stable gross margin delivery.