ASML and TSMC Increase Capacity Amid AI Chip 'Second Wave'

The AI semiconductor expansion cycle is far from over, but under high expectations, "strong demand" no longer automatically serves as a new catalyst for price increases.

Article by Jim, MSX MaiTong

Edited by: Frank, MSX Maotong

Last week, the pressured AI hardware sector waited for a strong catalyst to shift sentiment.

Subsequently, ASML and TSMC both delivered earnings reports with solid fundamentals that fell slightly short of elevated expectations: ASML significantly raised its full-year revenue and gross margin guidance and began increasing lithography machine capacity for 2027–2028; TSMC maintained its revenue, gross margin, and operating margin at historic highs while raising its full-year capital expenditure target outright to $60–64 billion.

Ideally, this should be the perfect combination for AI semiconductors—equipment manufacturers confirm that customers are still placing orders, while leading foundries demonstrate that orders are being converted into revenue and are willing to continue investing heavily in capacity expansion.

However, the market's response has not fully matched the strength of the performance.

The reason is not that the fundamentals of either company have deteriorated, but rather that expectations for the AI industry chain have been pushed to an unusually high level. The market is no longer satisfied with "strong demand"; it now expects every earnings report to be revised upward, every margin to break records, and every massive capital expenditure to immediately translate into higher profits.

This also conveys two signals from ASML and TSMC’s earnings reports—seemingly contradictory but simultaneously true: the expansion cycle for AI semiconductors is still ongoing, with some critical segments even accelerating; however, capital markets have shifted their pricing focus from validating demand to validating returns.

ASML was the first to reveal the answer to this earnings season.

The company's second-quarter net sales reached €9.326 billion, exceeding the previous guidance of €8.4 to €9.0 billion; the gross margin reached 54%, and net profit amounted to €2.918 billion. The company subsequently raised its third-quarter sales guidance to €11.0 to €12.0 billion and significantly increased its full-year 2026 sales forecast from €36.0 to €40.0 billion to €43.0 to €45.0 billion.

More important than quarterly data is ASML’s adjustment of equipment production capacity over the next two years. The company plans to increase low-NA EUV capacity by 30% from approximately 65 units in 2026 to 2027, and also raise DUV immersion equipment output by 30% from around 130 units. Meanwhile, ASML is exploring the possibility of further expanding production in 2028.

The lithography machine supply chain is complex, with long delivery cycles; ASML would not hastily increase production capacity for two years from now based solely on short-term order fluctuations. Such an expansion plan indicates that wafer fab customers are securing advanced process and high-end memory capacity for 2027–2028 in advance.

One day later, TSMC provided the corresponding verification from the wafer manufacturing side.

The company achieved revenue of $40.2 billion in the second quarter, representing a 12% sequential increase and reaching the upper end of its guidance range of $39 billion to $40.2 billion. Gross margin reached 67.7%, slightly above the upper end of guidance, while operating margin hit 60.3% for the first time. Net profit amounted to NT$706.56 billion, a 77.4% year-over-year increase, with earnings per share of NT$27.25.

The revenue mix continues to shift toward AI and advanced processes. In the second quarter, revenue from high-performance computing increased by 20% quarter-over-quarter, accounting for 66% of the company’s total revenue; advanced processes at 7 nanometers and below represented 77% of wafer revenue, with 3-nanometer and 5-nanometer technologies contributing 30% and 33%, respectively, while 2-nanometer, which has entered mass production ramp-up, contributed 3% of wafer revenue for the first time.

More significantly, capital expenditures remain the key indicator. TSMC has substantially increased its 2026 capital expenditure plan from the original range of $52 billion to $56 billion to a new range of $60 billion to $64 billion. Approximately 70% to 80% will be allocated to advanced processes, while 10% to 20% will be directed toward advanced packaging, testing, mask manufacturing, and other related areas.

The company also raised its full-year U.S. dollar revenue growth forecast from over 30% to slightly above 40%. Management noted that demand related to AI remains extremely strong, with continued positive signals from cloud service providers and downstream customers.

ASML is preparing to increase capacity for lithography equipment, while TSMC is expanding its wafer manufacturing and advanced packaging capabilities through higher capital expenditures.

Therefore, when both the leading device manufacturer and the world’s largest wafer foundry raise their future investments, at least one thing is clear: AI semiconductor capital expenditures have not entered a contraction phase—rather, the supply chain is still accelerating capacity preparation for demand over the coming years.

The issue is that the market is no longer just waiting for a "target achieved" earnings report.

Since TSMC releases its revenue data monthly, the $40.2 billion in second-quarter revenue was largely already priced in by the market; therefore, the true expectations gap prior to the earnings release lay in gross margin, third-quarter guidance, and the extent to which capital expenditures could be raised.

From this perspective, TSMC's second-quarter gross margin of 67.7% exceeded the upper end of the company’s previous guidance of 65.5% to 67.5%, but was roughly in line with the revised consensus market expectations, failing to meet some investors’ more aggressive forecasts of nearly 69% or higher.

For the third quarter, the company expects revenue of $44.6 billion to $45.8 billion, representing approximately a 12% sequential increase at the midpoint; however, the gross margin guidance has decreased to 65% to 67%, with a midpoint of approximately 66%.

The decline in gross margin does not indicate weakening demand.

TSMC expects the rapid ramp-up of its 2-nanometer process to dilute gross margin by approximately 3 to 4 percentage points in the second half of the year; expansion of overseas fabs will continue to increase depreciation and manufacturing costs. Strong demand for advanced processes, higher capacity utilization, and improved manufacturing efficiency can only partially offset these pressures.

In other words, TSMC is facing a classic boom-time expansion paradox: the stronger the demand, the more the company must procure equipment, build fabs, and ramp new processes in advance; the higher the capital expenditure, the sooner depreciation, overseas production costs, and the pressures of ramping new nodes will impact profitability.

This is also the most important aspect to understand in this financial report.

Based on management’s statements during the earnings call, TSMC does not seek to maximize short-term gross margins at the peak of supply tightness. The company emphasizes its role as a long-term partner to its customers, avoiding sudden, substantial price hikes that could pressure clients, and instead aims to maintain profit levels sufficient to support sustainable long-term expansion.

This means TSMC currently prioritizes balancing pricing power, customer relationships, and sustained capacity expansion over fully capturing all of its scarcity premium at once. From an industry perspective, this is undoubtedly a positive signal; however, from a short-term trading standpoint, it implies investors must accept the reality that while AI demand remains strong, not every additional dollar of revenue will immediately translate into higher profit margins.

Therefore, the market's muted reaction to TSMC's earnings report cannot be simply interpreted as AI demand having peaked; a more accurate explanation is that, in an environment where expectations were already exceptionally high, strong performance has become a necessary condition for valuation—but no longer automatically serves as a new catalyst for price increases.

After the earnings report, the strong performance did not immediately translate into sustained sector gains, reflecting investors' efforts to digest margin pressures and elevated expectations.

When viewing the earnings reports of ASML and TSMC together, the outline of the so-called "second wave" of AI chips has become clearer than before.

It is not a return to the widespread shortage where “all chips are insufficient,” nor is it a simple replication of the past two years’ market driven primarily by NVIDIA GPUs; rather, supply bottlenecks continue to spread across the entire AI system.

ASML’s EUV and DUV equipment determine how quickly advanced nodes can be scaled; TSMC’s 3nm and 2nm processes determine the wafer capacity available for GPUs, CPUs, and custom ASICs; HBM determines memory bandwidth; and advanced packaging technologies like CoWoS determine whether compute chips, memory, and high-speed interconnects can ultimately be integrated into deliverable data center products.

Any insufficiency in these steps will slow down the entire AI system's output.

TSMC management has explicitly stated that advanced packaging capacity is currently so constrained that it is limiting customer growth. The company is working to close the gap between demand and capacity, while welcoming additional packaging solutions to provide customers with more options.

Meanwhile, AI demand is expanding from single accelerators to a broader range of chip types.

TSMC believes that the development of agentic AI is re-elevating the importance of CPUs in data centers. Regardless of whether customers adopt x86, Arm, or RISC-V architectures, the advanced chips behind them still largely require TSMC’s manufacturing. This means that future AI capital expenditures will not only flow toward GPUs but will also continue to drive demand for CPUs, networking chips, memory, and advanced packaging.

Management also expressed optimism regarding long-term demand. TSMC believes that AI-related trends will remain strong through 2029–2030, with potential short-term fluctuations expected, but the long-term trajectory remains unchanged. Regarding its previous estimate of a 50% compound growth rate for AI-related business, management did not provide new specific figures, but indicated that demand trends are stronger than previously anticipated.

But this does not mean that all semiconductor companies will benefit equally.

• ASML (ASML.M) directly benefits from demand for lithography equipment and expansion of advanced process nodes;
• Applied Materials (AMAT.M), Lam Research (LRCX.M), and KLA (KLAC.M) benefit from demand for deposition, etching, and inspection equipment, respectively, but the pace of order fulfillment may vary;
• TSMC (TSM.M) controls advanced wafer manufacturing and packaging capabilities and is the primary beneficiary of AI chip production expansion.
• SK Hynix (SKHY.M), Micron (MU.M), and Samsung Electronics supply HBM and high-end memory;
• NVIDIA (NVDA.M), AMD (AMD.M), Broadcom (AVGO.M), and cloud providers with in-house chip development capabilities such as Amazon (AMZN.M), Alphabet (GOOGL.M), Microsoft (MSFT.M), and Meta (META.M) collectively determine how quickly end-demand can grow.

They are in the same capital expenditure cycle but have entirely different technological barriers, capacity constraints, profit structures, and valuation levels. Therefore, the "second wave" of AI chips is more likely to be a structural rally rather than a synchronized uptick across the entire hardware supply chain.

In the next phase, the market will pay closer attention to which companies truly control scarce, hard-to-replicate production capacity, which are merely following customers in increasing capital expenditures, and which can sustainably improve free cash flow and return on capital after expanding production.

Following ASML and TSMC, the next critical validation will fall to cloud providers such as Microsoft, Amazon, Google, and Meta. Ultimately, equipment manufacturers will be willing to expand production, and foundries will be willing to invest—but only if cloud providers continue to increase their capital expenditures and demonstrate that the growing AI infrastructure generates real model usage, enterprise revenue, and cash flow returns.

Objectively speaking, ASML confirmed that wafer fabs are still willing to purchase equipment, and TSMC further demonstrated that customer orders are sufficient to drive continued expansion of wafer and advanced packaging capacity.

From this perspective, the industry cycle for AI semiconductors has not yet peaked.

After all, with device production capacity expanding and advanced packaging still tight—even as TSMC raised its full-year capital expenditure to a maximum of $64 billion—these are not the signals of an industry preparing to contract.

But the market still feels unsatisfied because the issues in the next phase have shifted. Previously, investors needed to confirm whether AI demand was real; now, demand is hard to deny, and the market wants to know how much capital will be required to meet this demand, and how much profit and cash flow that capital can ultimately generate.

Therefore, the "second wave" of AI chips may have already begun, but it will not be a simple replay of the first rally.

What is truly scarce is no longer just companies that can produce more chips, but those that can control critical production capacity and, after massive expansion, continue to maintain pricing power, profit margins, and capital returns.