#ASMLTSMCQ3EarningsSoon

The AI Chip Bottleneck Is Moving Beyond Demand: ASML’s Equipment Capacity and TSMC’s Advanced Packaging Face the Next Test

The semiconductor industry is approaching an important checkpoint. Investors already know that AI companies need more computing power, but the next question is whether chipmakers can expand manufacturing capacity quickly enough to satisfy that demand without sacrificing margins or returns on investment.

ASML and TSMC occupy two different but connected positions in this equation. ASML supplies the advanced lithography systems used to manufacture leading-edge chips, while TSMC turns chip designs into physical silicon and provides advanced packaging technologies that help connect increasingly complex AI processors.

Their upcoming earnings could reveal where the next bottleneck is forming.

ASML: Can equipment capacity keep pace with customer demand?

ASML’s challenge is not simply selling more machines. It must expand its own production capabilities while customers plan increasingly ambitious semiconductor investments.

In its previous outlook, ASML said it intended to increase its 2026 low-NA EUV production capacity of approximately 65 systems by 30% in 2027. It was also investigating another potential 30% increase for 2028. For DUV immersion systems, the company planned a similar 30% capacity increase for 2027 from a 2026 level of approximately 130 systems. These are capacity plans, not guarantees of actual shipments.

The October 14 report should therefore be assessed through equipment deliveries, customer commitments, installed-base service revenue and order intake. If orders remain strong while ASML expands production, it would suggest that customers are preparing for sustained demand rather than making only short-term purchases. If new orders weaken, investors will need to question how quickly planned capacity can be absorbed.

TSMC: Advanced packaging may be as important as chip production

For TSMC, the key issue extends beyond the number of wafers manufactured. AI accelerators increasingly depend on advanced packaging technologies such as Chip-on-Wafer-on-Substrate, or CoWoS, to integrate processors and high-bandwidth memory into sophisticated packages.

Reuters reported ahead of the earnings release that demand for advanced packaging remained an important driver of TSMC’s business. Investors will be looking for evidence that the company can expand this capacity alongside demand for advanced manufacturing nodes.

I would focus on three questions during the October 15 earnings call: Is advanced packaging capacity expanding fast enough? Are customers maintaining their commitments for AI-related production? And can the company increase output without putting excessive pressure on margins?

A strong chip-demand environment can still produce bottlenecks if packaging, manufacturing equipment or other critical components cannot scale at the same speed.

The capital-spending question: Growth or overexpansion?

TSMC raised its 2026 capital-expenditure forecast in July to $60 billion–$64 billion, up from its previous forecast range of $52 billion–$56 billion. It is also investing heavily in U.S. manufacturing, including its planned Arizona facilities.

These investments demonstrate the scale of the capacity race, but they also create a financial test. More factories and advanced equipment require substantial upfront spending, and the returns depend on utilization, customer demand and future pricing.

Investors should distinguish between spending that supports contracted, visible demand and spending that assumes AI growth will remain exceptionally strong for years. Management’s explanation of utilization, capacity timing and investment returns could be just as important as the headline earnings numbers.

Three signals that could shape the market reaction

The constructive case would be supported by resilient equipment orders at ASML, clear progress in advanced packaging at TSMC and credible guidance showing that new capacity can be absorbed by customers.

The cautious case would emerge if order intake weakens, capacity expansion runs ahead of demand or management signals pressure on future margins and returns.

A third possibility is that both companies deliver strong operating results but their shares remain volatile because investors have already priced in substantial AI growth. Earnings beats alone do not guarantee positive stock-price reactions.

My focus for October 14–15

I would track ASML’s order momentum and capacity plans first, then compare TSMC’s advanced-node and packaging outlook with its capital-spending commitments. Together, these indicators offer a more useful view of the semiconductor investment cycle than looking at revenue growth in isolation.

The central question is no longer just whether AI needs more chips. It is whether the entire supply chain from lithography equipment to advanced manufacturing and packaging can expand efficiently enough to turn demand into sustainable profits.

That is the real test facing ASML and TSMC this earnings week.