Every advanced chip starts life under a beam of light. For years, almost the only machine that can draw the finest patterns has come from a single Dutch company. A Shanghai maker is now putting its own tools into Chinese fabs — and the question is no longer whether it can build one, but how good it can get.
Shanghai Micro Electronics Equipment (上海微电子装备, SMEE) is the name behind that effort. Founded in 2002 and tied to the Shanghai Electric group, it is the only Chinese company with the ability to develop and build front-end lithography scanners. Its progress is a useful window into how realistic "made-in-China" chipmaking really is.
The machine that draws the chip
Lithography is the step that prints a circuit pattern onto a silicon wafer using light. The finer the feature you can print, the more transistors fit on a chip. The global cutting edge belongs to ASML's EUV systems, which are off-limits to Chinese fabs under export controls. SMEE works a rung below that, in the deep-ultraviolet (DUV) range, where most of the world's chips are still made.
SMEE's 600-series scanners, built on 90nm-class ArF technology, are already in volume production and reportedly hold the large majority of China's domestic front-end lithography market. That is the mature-node business — the microcontrollers, power chips, and sensors that fill everyday electronics — and it is where Chinese fabs run today without foreign tools.
The 2025 step toward 28nm
The harder milestone is the 28nm-class immersion DUV system (model SSA/800-10W). According to industry and Hong Kong media reports in 2025, SMEE began delivering these machines to customers and moved into batch delivery through the year, with stated overlay accuracy in the low-nanometer range. Because SMEE does not publish audited spec sheets, treat those numbers as indicative rather than certified; the broader point — that a domestic immersion tool is now reaching 28nm-class work — is what multiple trackers agree on.
Why 28nm matters: it is the threshold below which a huge share of the world's "good enough" logic and many AI-adjacent chips are made. A domestic 28nm tool does not beat ASML's EUV, but it covers the bulk of real demand and, combined with multi-patterning tricks, can stretch even further.
The contract that signals "good enough"
In late December 2025, a filing on China's government procurement website showed SMEE winning a bid worth about 110 million yuan (≈ US$15.3 million) for an SSC800/10 stepper scanner — a 248nm KrF system used in advanced packaging, Mini/Micro LED, and MEMS. The purchase was handled as a single-source procurement by the Ministry of Science and Technology, a procedure reserved for cases where a supplier's technology is considered irreplaceable.
That detail is more telling than the price. A government body choosing SMEE as the only qualified source is a practical vote of confidence in domestic lithography for the packaging and specialty segments, even if it is far from the leading edge.
Where SMEE is already strong
Outside the front-end race, SMEE leads in areas the headlines ignore. Its advanced-packaging lithography tools reportedly hold a large share of the global market and an even larger share at home, and its LED lithography systems are among the most widely used in China. Packaging is where separately made chips are stacked and connected — exactly the chiplet and 2.5D work that Chinese compute vendors increasingly rely on.
A February 2025 spin-off, AMIES (芯上微装), was created to commercialize the equipment line faster, a sign the company is trying to turn prototypes into a steady supply business.
What readers can do now
- If you follow semiconductors, separate "lithography" into three buckets — mature-node (SMEE is viable), 28nm-class DUV (SMEE is entering), and EUV (still absent). Mixing them up causes both panic and hype.
- If you source hardware, expect more Chinese-made packaging and mature-node chips to carry SMEE-printed wafers; the impact shows up as supply resilience, not as faster phones.
- If you track the trade story, watch SMEE's component-localization rate and any new procurement wins; those are better signals than any single demo.
Honest limitations
This article draws on a government procurement filing (an official source) plus Hong Kong and Chinese technology-media reporting. SMEE does not publish independently audited yield, throughput, or overlay data, and some performance figures cited by media could not be cross-checked against a primary specification sheet. EUV-class capability is not claimed here and remains outside China's current reach. No investment advice or political judgment is intended.