ASML struggling to keep up with demand for chipmaking equipment as AI spending continues to boom

AI spending is pushing chipmakers to expand production far into the future. They need advanced manufacturing tools before they can make more processors, memory chips, and other semiconductors. Reserving ASML equipment early helps them plan future factories around a scarce and essential supply of machines.

ASML's customers include Samsung, Intel, and TSMC, which manufactures chips for Nvidia. SK Hynix and Micron are also expanding and need ASML machines. Their combined investment plans show that demand is not limited to one company or one type of chip.

The timing is striking because customers are already discussing output for 2029 and 2030. ASML plans to raise capacity by 30% in 2027 and another 30% in 2028, yet CEO Christophe Fouquet says demand may still surpass capacity. Early orders therefore reflect both strong AI expectations and concern that equipment will remain limited.

ASML makes the highly specialized equipment used to manufacture advanced semiconductors. It does not make processors itself. Instead, chipmakers use its machines during fabrication to create the microscopic patterns that become transistors and interconnections on silicon wafers.

In lithography, a machine projects or otherwise uses patterned light to mark selected areas of a wafer covered with light-sensitive material. After processing, those patterns guide later steps that build and remove material. Repeating the process creates increasingly complex circuits. This explanation uses established semiconductor knowledge beyond the article's detail.

The article identifies ASML as the largest manufacturer of tools for making AI chips and advanced semiconductors. Its customers include Samsung, Intel, TSMC, SK Hynix, and Micron. Because these companies are expanding production, demand for ASML's equipment is rising sharply, helping explain its €586 billion market valuation and planned price increases.

The article gives no exact figure for how much demand exceeds ASML's future capacity. It does establish a clear mismatch risk: customers are already seeking equipment for 2029 and 2030, while CEO Christophe Fouquet says demand could surpass capacity despite expansion.

ASML plans to increase capacity by 30% in 2027 and another 30% in 2028. If those percentages are applied successively, capacity would become about 1.69 times its starting level, or roughly 69% higher. That is a calculation, not an ASML forecast, and it does not show that demand will rise by the same amount.

The practical point is that even a very large expansion may not be enough. ASML also plans to raise prices starting in 2027, with the full effect in 2028. Strong orders, rising prices, and long-range reservations indicate that equipment availability remains a central constraint for AI-chip production.

Limited ASML equipment can restrict how quickly chipmakers add production. If new factories cannot obtain enough tools, supplies of advanced processors and memory could grow more slowly than planned. Strong demand combined with limited output can also keep chip prices high, although the article does not forecast a specific price increase.

ASML will raise prices starting in 2027, with the full effect appearing in 2028. The article already links higher component costs to AI spending: Microsoft CFO Amy Hood attributed $25 billion of Microsoft's capex budget to higher memory-chip and other component prices. This shows how equipment and component costs can feed into larger investment budgets.

AI data centers may therefore cost more to equip if chips become pricier or harder to obtain. That final effect is an implication, not a quantified claim in the article. Still, Gartner expects AI spending to reach $2.7 trillion in 2026, so even modest cost pressure could affect enormous construction and infrastructure budgets.

ASML sells equipment to major logic-chip manufacturers, including Samsung, Intel, and TSMC. It also supplies memory-chip makers SK Hynix and Micron. These companies use the machines to expand semiconductor production as AI demand increases.

TSMC has a particularly visible connection to Nvidia: the article says TSMC manufactures chips for Nvidia. Nvidia's AI systems therefore depend on a broader chain that includes chip design, fabrication, memory production, and the equipment used inside factories. ASML sits at the manufacturing-equipment stage of that chain.

The companies are investing because demand for AI chips and advanced semiconductors remains strong. The article also mentions Terafab, Elon Musk's planned semiconductor facility, which reportedly will receive its first ASML EUV tool in mid-2027. This suggests future demand could extend beyond today's established chipmakers and factories.

Chipmakers can combine different lithography technologies for different manufacturing steps. Older deep-ultraviolet systems can produce many layers, while the most advanced layers may require extreme-ultraviolet equipment. This distinction comes from established semiconductor knowledge; the article itself does not discuss alternative suppliers or technologies.

Nikon and Canon also make lithography equipment, especially for less advanced or specialized processes. Using these systems can reduce reliance on ASML for some layers or chips. However, switching is not simple because factories, process recipes, software, and production plans are designed around particular tools.

Alternative equipment therefore may ease pressure in selected areas, but it is unlikely to remove dependence on ASML for the most advanced production. The article highlights ASML as the largest manufacturer of tools for AI chips and advanced semiconductors. Its customers' future orders and ASML's own capacity expansion show how difficult that bottleneck is to avoid.

Semiconductor lithography is a pattern-transfer process. A wafer is coated with a light-sensitive material, and a lithography machine projects a circuit pattern onto selected areas. The exposed material changes chemically, creating a temporary map for later manufacturing steps. This technical explanation uses established semiconductor knowledge beyond the article.

The wafer then undergoes processes that remove or add material according to the mapped areas. Engineers repeat exposure and processing many times, using different patterns for different layers. Those layers form transistors, wiring, and other circuit structures. Extremely precise alignment is essential because modern processors contain huge numbers of tiny features.

ASML makes the advanced tools that enable this work, including EUV systems mentioned in the article. Its customers use the machines to make AI chips and advanced semiconductors. Because each factory needs specialized equipment and AI demand is expanding, ASML's limited capacity can influence how quickly new processor production comes online.