TrendForce, citing Chinese semiconductor outlet ijiwei and Electronic Times, says conventional etch and thin-film deposition tools are now taking about 12 months, while some advanced packaging and test equipment exceeds 18 months. That is enough to change fab schedules. Equipment is ordered far ahead of production, installed in stages, qualified on a process line, and then tuned for yield. A delayed delivery is therefore more than a late capital purchase: it can move the point at which a new memory or logic line begins producing saleable wafers.
For Windows PC buyers and the IT organizations planning around AI PCs, servers, and storage, the immediate effect is unlikely to be a repeat of the retail GPU shortages of 2021. The risk is further upstream. Persistent bottlenecks in the tools used to build DRAM, NAND, and advanced logic can prolong pressure on the supply of high-bandwidth memory, enterprise SSDs, accelerators, and eventually the processors that reach PCs and data centers.
The bottleneck has moved upstream
The submitted report is directionally supported by both vendor disclosures and independent reporting. Electronic Times found that equipment suppliers serving TSMC and Micron had seen some delivery windows rise from three or four months to six to eight months, while another packaging-equipment supplier reported some orders now taking more than a year. The outlet said Samsung Electronics and SK hynix were reviewing earlier ordering to avoid having equipment availability dictate fab start dates.
This is a different kind of semiconductor constraint from a shortage of finished chips. A chipmaker cannot simply substitute a missing deposition chamber, etch system, metrology tool, or RF power supply with another component from inventory. These tools are part of an integrated manufacturing flow, and qualification can be specific to a process, a fab, and sometimes an individual production line.
The reported 12-to-24-month figures should therefore be treated as a range of supply-chain pain points, rather than a universal lead time across the equipment industry. The independently reported evidence supports roughly year-long delays for major tools and significant extensions at suppliers. TrendForce’s 18-month packaging-and-test estimate and its 24-month RF-power-supply estimate are more specific claims attributed through ijiwei; no comparable public vendor-wide dataset was identified to establish them as an industry average.
That distinction matters because it changes the conclusion. The industry is not out of every kind of chipmaking equipment. It is facing a capacity-planning problem in which certain equipment categories and supporting subsystems can become the pacing item for a multibillion-dollar fab project.
ASML’s order book confirms demand, not every lead-time claim
ASML’s July 15 second-quarter results provide the clearest public confirmation that demand for major equipment is running well ahead of normal planning horizons. The company said it would add 30% to its 2026 Low-NA EUV production capacity of around 65 systems for 2027 and was investigating another 30% expansion in 2028. It announced the same 30% 2027 increase, and a possible further increase in 2028, for DUV immersion capacity starting from roughly 130 systems in 2026.
ASML CFO Roger Dassen said during the company’s investor call that the company was close to having the EUV orders needed for 2027, even with the planned capacity increase, and had already received a large number of orders for 2028. Reuters and other financial reporting independently described the 2027 EUV capacity as close to fully booked.
That is strong evidence of a real tool-capacity squeeze, but it should not be used to validate every number in TrendForce’s report. ASML sells lithography systems; the cited shortages span etch, deposition, packaging, test equipment, and RF power infrastructure. The same investment surge can stress all of those layers, but they have different suppliers, lead times, replacement options, and technical barriers.
SEMI’s published forecast supports the broader spending backdrop. It projected global semiconductor manufacturing-equipment sales would reach $145 billion in 2026 and $156 billion in 2027, driven by AI-related leading-edge logic, memory, and advanced packaging investment. Its forecast specifically identified growth in DRAM equipment spending as memory makers expand high-bandwidth-memory capacity, and it expects NAND equipment spending to rise as suppliers add capacity and increase 3D NAND layer counts.
The equipment constraint is thus being driven by several overlapping buildouts: leading-edge logic for AI accelerators, HBM and advanced DRAM, denser 3D NAND, advanced packaging, and mature-node factories supported by industrial-policy programs. No single PC component is responsible, and no single toolmaker can fix it alone.
China’s opportunity is real, but it is concentrated
TrendForce’s central argument is that Chinese equipment makers can gain share when overseas tools take longer to arrive. That conclusion is plausible, particularly in process steps where local suppliers already have commercial products and foundries have an urgent need to expand capacity.
The report says AMEC can deliver etch systems in about nine months and Piotech can deliver certain CVD thin-film deposition tools in roughly 10 months, compared with reported 12-month waits for conventional imported tools. Those delivery estimates are attributed to ijiwei; neither company has publicly supplied a standard lead-time schedule that would allow a direct, apples-to-apples comparison across configurations and customers.
Still, domestic Chinese toolmakers are no longer theoretical competitors. Reporting cited by Tom’s Hardware from Nikkei Asia found that Naura, AMEC, and Piotech posted steep revenue growth through 2025 as Chinese fabs bought more locally sourced equipment. But the same reporting included an important counterweight: domestic vendors are competing intensely on price, and analysts have flagged deteriorating margins at some firms.
That means a faster quoted delivery date does not automatically equal a durable market-share win. A tool has to meet process requirements, integrate with existing fab automation, pass customer qualification, sustain uptime, and produce acceptable yields. For mature nodes, some memory steps, and less technically demanding process layers, a domestic alternative may be increasingly attractive when the imported equivalent is a year away. For the most demanding production steps, qualification and performance remain more important than a few months of purchasing lead time.
The reported opening for Chinese vendors is therefore greatest in the segments where they are already established—etch, deposition, cleaning, certain packaging and test categories—not across the entire equipment stack. ASML’s EUV systems are the obvious counterexample: its expanding order book demonstrates how difficult it remains to substitute the most advanced lithography capacity.
Earlier orders will become a competitive advantage
The practical response from chipmakers is already emerging: order earlier, reserve capacity sooner, and treat equipment availability as part of product planning rather than as a procurement detail. Electronic Times reported that Samsung and SK hynix were considering earlier orders as they pursue multiple fab projects. The companies that secured supplier slots before the demand surge have an advantage that cannot be quickly replicated by competitors with available capital but no place in an equipment queue.
That can reinforce the market power of the largest chipmakers and cloud customers. A memory maker building HBM capacity for contracted AI customers can justify early tool orders and large deposits. A smaller company or a fab targeting a more uncertain market may hesitate, then find itself waiting behind the earlier commitments.
For PC manufacturers, the impact will arrive indirectly. Windows laptops and desktops depend on a supply chain that competes with AI servers for leading-edge logic, DRAM, NAND, packaging capacity, and increasingly the equipment needed to expand all of them. The strongest near-term pressure remains in data-center and AI components, where buyers can absorb higher prices. But if equipment delivery constraints persist into 2027, those high-margin priorities can slow the normal easing of memory and storage costs that consumer-PC buyers usually expect after a capacity expansion cycle.
The missing number is how much supply can be added
The most consequential unanswered question is not whether lead times have risen—they have—but how quickly equipment manufacturers and their own suppliers can add usable output. ASML has put a public number on its plan: 30% more Low-NA EUV and DUV immersion capacity in 2027, with another possible 30% increase in 2028. Most other major toolmakers have not publicly supplied equivalent category-by-category capacity figures that would show when the reported bottlenecks in etch, deposition, packaging, test, and RF subsystems will ease.
Until those supply additions arrive, semiconductor companies will compete for delivery slots as aggressively as they compete for wafers. China’s local equipment makers may capture meaningful orders where their systems are qualified and immediately available, but the broader shortage is more likely to reward suppliers with proven capacity than to erase the technical gap between domestic and global leaders.