Since December 2025, a wave of reporting originating with a Reuters investigation and amplified across trade press and financial media has described a classified Chinese state program to build an indigenous Extreme Ultraviolet (EUV) lithography capability. Sources close to the effort described Chinese scientists building, in a high-security Shenzhen laboratory, a prototype machine capable of producing the advanced chips that power AI systems, smartphones, and military hardware technology previously monopolized by the West. The program has been informally nicknamed China's "Manhattan Project" by people familiar with it, evoking the wartime US effort to build the atomic bomb.
Before assessing what this means, it's worth separating three things the reporting supports from one figure that doesn't appear to be well-sourced.
The $40 billion figure doesn't check out as stated
The number circulating in some summaries "$40 billion funneled into the EUV program" isn't the figure that shows up in the underlying reporting. What the record actually shows:
●China's "Big Fund" Phase III, the state semiconductor investment vehicle, is being redirected toward lithography and EDA with roughly $47.5 billion in capital but that fund covers China's chip industry broadly, not the EUV program specifically.
●China's overall spending on wafer fab equipment (all tool categories, not just lithography) hit a record $49.5 billion in 2024, and China remains the world's largest buyer of chipmaking tools through at least 2027.
●Separately, one Reuters-linked account puts the timeline at roughly years of state subsidies poured into the project without a hard aggregate dollar figure.
No source found attributes a specific $40 billion figure to the EUV lithography program itself. This matters: conflating national semiconductor-fund totals with a single black-budget project overstates precision that the public reporting doesn't actually have. Treat any specific dollar figure attached to this program with caution it's plausible the initiative costs tens of billions given the scale described, but no on the record figure that precise has surfaced.
What is well-documented
The prototype is real, but limited. Reuters reported the machine was completed in early 2025, fills nearly an entire factory floor, and is operational — successfully generating extreme ultraviolet light — but has not yet produced a working chip. Independent trade analysis corroborates this: the prototype generates EUV photons but hasn't yet exposed a wafer.
Scale of the effort. Reporting describes a program involving more than 3,000 researchers, with a dedicated team of roughly 100 recent university graduates focused specifically on reverse-engineering components from EUV and DUV machines, reportedly obtained via secondary markets older ASML equipment has been resold through Chinese auction platforms as recently as October 2025.
Former ASML personnel. Multiple accounts describe the technical breakthrough as substantially dependent on ASML veterans, including recruitment of Western scientists, a pattern Dutch intelligence separately flagged, warning that China "used extensive espionage programmes in its attempts to obtain advanced technology and knowledge from Western countries." Reuters said it could not determine whether legal action has been taken against any former ASML employees involved.
Huawei's coordinating role. This is the most consistently reported element. Huawei is described as involved in every step of the supply chain — from chip design and fabrication equipment to manufacturing and final integration into products like smartphones while formally the project remains state-run. One account adds that Huawei CEO Ren Zhengfei personally briefs senior Chinese leaders on the program's progress.
The timeline is aggressive and slipping. The government's internal target was producing working chips on the prototype by 2028, though sources cited by Sherwood News indicate China is likely a couple of years behind that schedule, pointing to a more realistic production timeline around 2030.
Why lithography is the actual bottleneck
This is the part that skeptics of the "China caught up" framing should sit with. Even amid the EUV headlines, China's broader tool localization numbers show the gap is concentrated exactly where you'd expect:
On raw output lithography toolmakers are stuck producing roughly five machines a year, against ASML's 130.
In other words: China has made real, non-trivial gains in most of the semiconductor equipment stack. Lithography and EUV specifically remains the one place where the "just throw state money at it" strategy has yet to close the gap at any meaningful volume. A single working prototype, even a genuine one, is a long way from a production-line tool that can compete with ASML's decades of accumulated precision optics and systems-integration know-how.
The strategic read
For Beijing: the prototype is a proof-of-concept that justifies continued investment and gives domestic champions (SMIC, Naura, SMEE) political cover to consolidate — reporting notes a March 2025 call from SMIC's co-founder and other industry heads for national consolidation of what they described as a "small, fragmented, and weak" tool industry.
For Huawei: deep integration across the chip supply chain — from design to fab tools to final assembly positions the company as the likely first beneficiary if the technology scales, particularly for its Ascend AI accelerators, which face the most direct pressure from ongoing US export controls.
For Washington and the Netherlands: the reporting is a signal that export controls have slowed China's advanced-chip trajectory but not stopped it, and that human capital (recruited engineers) and secondary markets (resold used equipment) are harder to control than direct equipment sales. ASML has stated no EUV system has ever been sold to a customer in China, and the company says it restricts EUV knowledge internally to a limited set of employees — yet the prototype exists anyway.
For the "AI chip war" narrative broadly: It's worth resisting the framing that this settles the race. A working light source in a lab is a necessary but nowhere-near-sufficient condition for volume chip production. The precision optics, overlay accuracy, and systems-integration problems that make EUV hard are not solved by generating EUV photons — they're solved by hitting sub-nanometer accuracy at high throughput, repeatably, for years, which is the part ASML has never lost.
Bottom line
The underlying facts a real, Huawei-coordinated, state-funded prototype EUV light source built partly with reverse-engineered ASML parts and former ASML engineers — are corroborated across multiple independent outlets tracing back to the original Reuters investigation. The "$40 billion" figure attached to this specific program, however, doesn't appear in that reporting and shouldn't be treated as confirmed; the verifiable dollar figures (Big Fund III's ~$47.5B, $49.5B in 2024 tool spending) describe China's semiconductor push broadly, not this program in isolation. The realistic takeaway is a partial, contested breakthrough with a multi-year runway to production — not a fait accompli that neutralizes existing export controls.
