2026-09-01 — UBS estimates China needs 2-5 years for immersion-DUV high-volume manufacturing, a decade-plus for any EUV alternative

• ⚠️ OUT OF WINDOW (2026-09-01, before this sweep’s 2026-09-02 15:00 ET start) — flagged because it directly answers this thread’s central open question and is not recorded anywhere on the map. UBS analysts, cited by TrendForce and reported separately, estimate China will reach immersion-DUV high-volume manufacturing capability within two to five years, based on patent activity around light-source and laser subsystems they read as comparable to where ASML stood roughly 15 years before its own EUV mass-production ramp. On EUV specifically, UBS places China’s program roughly where ASML was in 2004 and does not expect a viable EUV alternative to emerge within this decade. This is an analyst timeline forecast, not a delivery confirmation — it does not meet this thread’s own “delivery, not production” test (no installed-and-running tool at a named fab is newly confirmed here) — but it is the first dated, sourced numeric range this map has seen for the gap this thread was opened to track. (TheNextWeb, ChipBriefing)

2026-08-29 — Goldman puts a number on the gap this thread exists to test, and names lithography as the reason it doesn’t close faster

• Goldman Sachs forecasts China’s advanced-chip deficit narrowing from 92% short of demand today to 34% short by 2035 — i.e. roughly two-thirds domestic self-sufficiency in advanced chips within a decade, per a research note relayed by SCMP (08-24) and corroborated by three independent outlets (Tech Times, a Korean business-press pickup) over 08-24/26 with consistent figures. The explicit caveat matches this thread’s own test exactly: SMIC’s manufacturing yield and lithography capability are named as the binding constraint on how fast the gap closes — one Tech Times headline states it as “Lithography Still Binding Limit.” This is the first outside quantification found of the ceiling this thread has been tracking qualitatively since 07-28; it does not report any new delivery event and the DELIVERY test below is still unmet. ⚠️ Could not reach Goldman’s underlying note or the SCMP original directly (paywall/fetch-blocked); figures cross-checked across four independent headlines rather than a single source. (SCMP, “China’s advanced chip deficit to shrink from 92% to 34% by 2035, Goldman says” · Tech Times, “Goldman Sees China 34% Short of Chip Self-Sufficiency in 2035; SMIC Yield Is Fragile Key” — via Google News RSS)

2026-08-18 — Three weeks on, still nobody has confirmed a delivered tool

Logged 2026-08-18. This thread’s own test is DELIVERY, NOT PRODUCTION — “a tool installed and running at a named fab, at a stated node and yield.” That test is still unmet, and the non-result is the finding.

• No confirmation exists, from Shanghai Aishengna or from any of SMIC, Hua Hong or CXMT, of an installed, running tool at any node with any yield or uptime figure. Twenty-one days after the Reuters report. The thread said “a slip here is as informative as a hit”; this is the slip, and the 2026 target of ~5 systems still has four months to run.
• A hard qualifier this map did not have: the tools support 28nm-class single-exposure patterning only. More advanced nodes would require multi-patterning, and some critical components remain internationally sourced with local-supplier delays. That materially narrows what “China has domestic immersion DUV” means — and this thread is load-bearing under the whole chokepoint picture, so the narrowing propagates.
• The market repriced and the company did not. ASML fell as much as 14% on the 07-27/28 report — the largest single-day reaction on any thread in this rotation, with Samsung, SK Hynix, Micron, SanDisk and AMD also selling off. Then ASML’s own Q2 raised 2026 net-sales guidance to €36–40B from €34–39B, reaccelerated DUV immersion shipment guidance to ~130 systems (roughly flat year on year), and described the order book as “largely filled through 2027,” with China still ~20% of net sales. No export-control policy change was found tied to this story.
• ⚠️ Naming collision resolved: “Shanghai Yuliansheng”/“Yuliangsheng” and “Shanghai Aishengna” are not rival efforts — the former is a predecessor team absorbed into the latter, alongside the pre-existing state champion SMEE. Two same-day 07-28 reports used different names without cross-referencing, which is exactly how a single programme gets double-counted as a race.

2026-07-30 — ASML’s slide gets a number; delivery still hasn’t fired

• ASML’s slide is now quantified: -7%, building into a “~€60B wipeout” framing across outlets — pricing the same 07-28 announcement, not a new fact. SMIC and smaller China-chip names also down. ([Seoul Economic Daily])
• The thread’s own test — delivery, not announcement — still has not fired. No named fab confirmed receiving a tool.
• ⚠️ Unconfirmed, not used: a Digitimes claim of “$1B state backing” for Aishengna is paywalled with no independent corroboration.

2026-07-29 — Split out on Ben’s steer; Shanghai Aishengna named as the maker

• Split out of China’s Stack on Ben’s steer — “Sure, add it. its interesting. Probably a sub-thread of china-stack-independence.” The parent was promoted story → meta to take the pointer.
• Shanghai Aishengna was named as the maker — founded August 2023, backed by Shanghai Electric Holding and Shanghai International Trust, staffed from SMEE and Yuliangsheng. (Seoul Economic Daily)
• JPMorgan’s caveat: “producing a small quantity of DUV equipment and building a mass-production system are entirely different matters.”
• No delivery to any named fab has been confirmed — coverage still says “expected this year” for SMIC, Hua Hong and CXMT.
• The thread’s test is delivery, not announcement, and it has not fired.

2026-07-28 — China starts mass-producing its own immersion DUV

• Reuters names the maker: Shanghai Aishengna Electronic Technology Group, state-owned, assembled by absorbing lithography-startup teams — China’s first mass-produced immersion DUV tool. ~5 units targeted for 2026, ~20 for 2027, with SMIC, Hua Hong and CXMT the expected first recipients this year. (Reuters)
• The caveat is half the story — the tools are “far from matching” ASML’s competing models and are unproven at high-volume manufacturing. A first rung, not parity. What makes it trackable is that it comes with numbers: a maker, a unit count, and named recipients. Ledger entry china-duv-units-2026 opened on the 2026 target.
• It priced immediately — cited across the tape as a catalyst for the fourth straight losing session in the chip complex and for Korea’s second circuit breaker in two days; one sell-side note framed it as a possible “DeepSeek 2.0 moment” for semis.

← Backstory