Chinese semiconductor thread II

tokenanalyst

Lieutenant General
Registered Member
This may well be the most important news of the last 2 years.

Because of this I am wondering about the sources: "According to reports" what it means according to reports?

They say "In March 2026, Shanghai Microelectronics publicly showcased the machine for the first time at SEMICON China 2026" but this is not what has been reported in this forum in March, actually the stand of SMEE was showing a mock-up that they present every year and I don't remember of any official announcement at the time: SEMICON China 2026 was covered at length here.

Also overlay tolerance ±2.3~2.5nm is the single most important parameter for a litho machine. It would be important to have an official spec by SMEE in this regards. The article seems to hint that the SSA800 machine is now publicly acknowledged and officially reported by SMEE: so why specs arrive through "reports" ?

Again, I am questioning not to minimize the news, but actually for the opposite reason, because the news if official (currently it is not IMHO) would be the single most important of the last few years: there is a big difference between knowing that the SSA800 exists and it works and its official presentation with official specs by SMEE.
Is not being sold in the open market for anyone to buy. Is going mainly to sanctioned companies who ASML doesn´t do business anymore. ASML has another things to worry about the Chinese market.

-Their revenue from the country is falling faster than others foreign companies.

- A potential full ban from the country that the company can´t avoid because rely in critical components made in the US like ASML Wilston for metrology and wafer stages and ASML Cymer for light sources.

-SMEE ( or whatever they name themselves now) KrF lithography machines are a bigger threat than the Immersion machines because those are being sold outside the sanction regime.

- More domestic lithography companies in the market, with some already producing 350nm I-line steppers, so is a matter of time that they start producing 250nm I-line scanners.

-Domestic litho Metrology and computational lithography companies like DJEL are already eating through that market share.

-The inability to sell EUV scanners is accelerating the developing of a domestic EUV lithography supply chain in China that could end up not being compatible with ASML EUV offerings IF there is some relaxation in the sanctions.
 

sunnymaxi

Colonel
Registered Member
Is not being sold in the open market for anyone to buy. Is going mainly to sanctioned companies who ASML doesn´t do business anymore. ASML has another things to worry about the Chinese market.
as per this report they will produce 5-8 units in Q2. it means full year production maybe like 20-30 units which is enough to supply many fabs in mainland.

remember, YMTC have already fully domestic line running. Beijing Yandong too. there are many mature fabs in mainland. Huahong transition to FinFet.
 

tphuang

General
Staff member
Super Moderator
VIP Professional
Registered Member
the overlay requirements for 7nm is 2.0nm so is very close.
well, it's just needed for the most critical layer IIRC?

so basically, this can fully replace 1980i IIRC (although at lower process rate).

-SMEE ( or whatever they name themselves now) KrF lithography machines are a bigger threat than the Immersion machines because those are being sold outside the sanction regime.
Where are you getting this? It's in fact likely that SMEE just concentrate on ramping Arf/Arfi scanner production because their WPH rate is likely lower than that of ASML machines. So, they would concentrate on delivering to top end labs.

This may well be the most important news of the last 2 years.

Because of this I am wondering about the sources: "According to reports" what it means according to reports?

They say "In March 2026, Shanghai Microelectronics publicly showcased the machine for the first time at SEMICON China 2026" but this is not what has been reported in this forum in March, actually the stand of SMEE was showing a mock-up that they present every year and I don't remember of any official announcement at the time: SEMICON China 2026 was covered at length here.

Also overlay tolerance ±2.3~2.5nm is the single most important parameter for a litho machine. It would be important to have an official spec by SMEE in this regards. The article seems to hint that the SSA800 machine is now publicly acknowledged and officially reported by SMEE: so why specs arrive through "reports" ?

Again, I am questioning not to minimize the news, but actually for the opposite reason, because the news if official (currently it is not IMHO) would be the single most important of the last few years: there is a big difference between knowing that the SSA800 exists and it works and its official presentation with official specs by SMEE.
if you read the article. It says the machine being used is already at 90-95% yield. Which is quite significant.
And that they are delivering 5-8 in Q2
150wpm is a good start but this has a long way to be improved! Overtime, the gap will shrink
I guess you can compensate with multiple machines? SSA800 is likely less economical than ASML machines to start off. But given the latter might not have maintenance, seems like worth the trade off.
 

tokenanalyst

Lieutenant General
Registered Member
Where are you getting this? It's in fact likely that SMEE just concentrate on ramping Arf/Arfi scanner production because their WPH rate is likely lower than that of ASML machines. So, they would concentrate on delivering to top end labs.
KrF development is pretty mature right now and it was pretty mature in the 600 series, RSLaser KrF light source was already developed in the 600 series, KrF optics and light sources are way cheaper and easier to make so I don´t think will bog down ArF/ArFi machines production .

Their machines where developed to be used along ASML machines, especially in metal layers and I think will be a prime target for entering the domestic lithography market for companies outside the sanction regime instead of going for the more risky ArF machines. Also can be used by mature nodes 130nm nodes by companies, like auto chips and MEMs.

Also you have to take into account that ASML makes machine for a global market and multiple pretty big companies. Domestic lithography machines performance has to be tied to domestic demand, for now, going for ASML performance from the start would probably bog down the development of domestic lithography machines unnecessary. The goal is not to compete with ASML, for now, but to meet domestic demands already sanctioned companies.
 
Last edited:

sunnymaxi

Colonel
Registered Member
KrF development is pretty mature right now and it was pretty mature in the 600 series, RSLaser KrF light source was already developed in the 600 series, KrF optics and light sources are way cheaper and easier to make so I don´t think will bog down ArF/ArFi machines production .

Their machines where developed to be used along ASML machines, especially in metal layers and I think will be a prime target for entering the domestic lithography market for companies outside the sanction regime instead of going for the more risky ArF machines. Also can be used by mature nodes 130nm nodes by companies, like auto chips and MEMs.
600 series is already in serial production for quite some time.

800 series specifically developed to compete in immersion category coz US always threaten to cut off ASML supply. article says 800 series achieved mass production. it means SMEE planning to compete with ASML in DUVi category and obviously can't compare economically with ASML initially but can compensate in numbers.

remember @interestedseal did post about Lithography suppliers mass production plan in 2026. i believe from 2026 onward SMEE will increase immersion scanner production further. maybe we see improved SSA800 variant soon. i m sure improved version is already under development or near completion. this is the Chinese pattern in pretty much every high tech field.
 

tokenanalyst

Lieutenant General
Registered Member
I guess you can compensate with multiple machines? SSA800 is likely less economical than ASML machines to start off. But given the latter might not have maintenance, seems like worth the trade off.
In my opinion. Due its monopoly status ASML operate under pretty high margins. SMEE? will operate under lower margins, for now. My guess that most components are domestically manufactured. Looks like production is already ramping up so will be more costs reductions than in the R&D phase and with a little bit of help from subsidies these machines can be procured an at ever decreasing cost.
 
Top