US Startup Substrate Announces Chipmaking Tool That It Says Will Rival ASML (reuters.com) 38
An anonymous reader quotes a report from Reuters: Substrate, a small U.S. startup, said on Tuesday that it had developed a chipmaking tool capable of competing with the most advanced lithography equipment made by Dutch firm ASML. Substrate's tool is the first step in the startup's ambitious plan to build a U.S.-based contract chip-manufacturing business that would compete with Taiwan's TSMC in making the most advanced AI chips, its CEO James Proud told Reuters in an interview. Proud wants to slash the cost of chipmaking by producing the tools needed much more cheaply than rivals. [...]
An engineering feat that has eluded even large companies, lithography needs extreme precision. ASML is the only company in the world that has been able to make at scale the complex tools that use extreme ultraviolet (EUV) to produce patterns on silicon wafer at a high rate of throughput. Substrate said that it has developed a version of lithography that uses X-ray light and is capable of printing features at resolutions that are comparable to the most advanced chipmaking tools made by ASML that cost more than $400 million apiece. The company said it has conducted demonstrations at U.S. National Laboratories and at its facilities in San Francisco. The company provided high resolution images that demonstrate the Substrate tool's capabilities. "This is an opportunity for the U.S. to recapture this market with a homegrown company," Oak Ridge National Laboratory director Stephen Streiffer, an expert on high-energy x-ray beams, said in an interview. "It's a nationally important effort and they know what they're doing."
An engineering feat that has eluded even large companies, lithography needs extreme precision. ASML is the only company in the world that has been able to make at scale the complex tools that use extreme ultraviolet (EUV) to produce patterns on silicon wafer at a high rate of throughput. Substrate said that it has developed a version of lithography that uses X-ray light and is capable of printing features at resolutions that are comparable to the most advanced chipmaking tools made by ASML that cost more than $400 million apiece. The company said it has conducted demonstrations at U.S. National Laboratories and at its facilities in San Francisco. The company provided high resolution images that demonstrate the Substrate tool's capabilities. "This is an opportunity for the U.S. to recapture this market with a homegrown company," Oak Ridge National Laboratory director Stephen Streiffer, an expert on high-energy x-ray beams, said in an interview. "It's a nationally important effort and they know what they're doing."
Snake oil for sale (Score:1, Insightful)
Invest your hard earned money while you can!
Re: (Score:3, Funny)
I'm surprised they didn't include "AI-powered" in the announcement.
Re: (Score:2)
Re: (Score:2)
I was looking out for quantum, which is a proven method for getting peoples brains to close down and their wallets to open up.
The best way to get people to throw money at you is to somehow sneak in a "AI powered by quantum blockchain." Just writing that feels like venture capitalism having an orgasm.
Re: (Score:2)
Re: (Score:2)
Would it be a quantum blockchain or quantum AI though? Actually since it's gibberish all the way down it could be a superposition of both.
Quantum AI on the quantum blockchain running on quantum computers running on quantum fusion reactors!
Never a good sign (Score:3)
Re: (Score:2)
Depends on the x-ray energy band and required power to achieve what is needed, but I suspect a copper or lead mask would suffice. If I knew their target energy level I could give a better idea of what would work for a mask.
Re: (Score:3)
Re: (Score:2)
Various fucking NATION STATES have been contemplating this for decades.
And now it turns out a startup in america will do it with VC money.
What China (and before it, Japan) have tried with infinite government money, and failed, a startup will do it with VC money.
Re: (Score:2)
VC's will seed it and buy the politicians and thus government contracts will fund it. The direct gains will be privatized and direct losses will be socialized.
Same way SpaceX spanked NASA and ULA. Just competent management and the drive to build rather than punch a clock.
As much as I dislike this form of National Socialism on principle, the best case a State can make is that it needs the capacity for sovereign production of defensive weaponry.
A State is sovereign to the degree that it is not reliant on ot
Re: (Score:2)
And now it turns out a startup in america will do it with VC money.
*may do it
I don't assume they will fail, but I also don't assume they will succeed.
There's no basis on which you should either.
Re: (Score:2)
Re:Never a good sign (Score:5, Informative)
From https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fsiliconangle.com%2F2025%2F10%2F28%2Fsubstrate-raises-100m-impossible-reinvent-chipmaking-industry%2F [siliconangle.com],
Substrate has already demonstrated its technology at the U.S. National Laboratories and at its own facilities in San Francisco, and offered some high-resolution images that show what it can do.
Experts told the Journal that the images are impressive, but questions remain about Substrate’s ability to print at the scale required for the mass production of chips. Indeed, many are extremely dubious about the company’s plans, noting that it would need to maintain the same accuracy shown in the images over much larger areas of the silicon wafers and run at incredibly fast speeds. Chinese companies have been trying for years to match ASML’s capabilities, but have not yet succeeded despite spending billions of dollars on research.
Re: (Score:1)
And ASML did not get to where they are at without buying US tech companies:
"ASML has acquired several US companies, including Cymer (2013) and Silicon Valley Group (SVG) (2001), as well as Brion Technologies in 2007. These acquisitions were strategic for obtaining critical technology, particularly light sources and lithography systems, to advance semiconductor manufacturing."
Re: (Score:2)
From the article, " Reuters was unable to independently verify the company's claims about its technology." That said, sounds like they plan on using X ray wavelengths, which in theory could provide much better resolution than even EUV. Problems will be many including finding a photo resist that works with Xray and chip manufacturing as just the first one.
So, for clarity, does this mean that this company is ahead or behind the Chinese companies that are also imminently catching up to ASML?
They are probably just lying (Score:4, Insightful)
Standard start-up behavior.
Re: (Score:3)
They could have had a grad student with access to the now decades old proximity x-ray lithography equipment at US National labs make a pattern on a wafer. Then it's not lying, exactly.
Of course that old equipment worked at very low power and proximity lithography is almost unusable. For high throughput you'd need a Billion dollar accelerator for your x-ray source.
Re: (Score:2)
Re: (Score:2)
Ok so... one of these might exist. But the desktop is at least 15 meters long and the laser is 0.3 PW. The main body of the system is a tube of plasma, so enjoy keeping that stable on a production line...
If you handwave that the laser is likely half a billion by itself then sure, this tech is... at the point where someone could start a serious development process to turn it into an industrial system. In 5 to 10 years, I'd estimate...
Looks tabletop to me... (Score:2)
Re: (Score:2)
EUV lithography light source power is sustained, this is 80fs pulses at 1kHz.
Re: (Score:2)
Re: (Score:2)
The average power of the EUV light source is about the peak power of this desktop xray source.
Re: (Score:2)
Re: (Score:2)
For x-rays you'd probably use diffractive optics. With sufficient coherence, the mask could be a hologram even.
Re: (Score:3)
No, you have it all wrong. Start-ups forecast the truth ... it will eventually happen ... trust us.
Re: (Score:2)
Indeed.
If this is true (Score:2)
If this is true then it won't be long until they sell the tech to the highest bidder of a foreign company. "This is an opportunity for the U.S. to recapture this market with a homegrown company" - grandiose idea until greed kicks in.
Re: (Score:2)
The US won't even permit companies in foreign countries from selling related tech willy nilly. Wtf make you think a domestic company with breakthrough lithography technology will just sell out to foreign randos however they please.
Try making some sense when you post stuff.
Re: (Score:2)
Sorry dad.
Prediction (Score:2)
If Dutch ASML feels threatened by the USA, all the do-not-sell-EUV-technology-to-countries-the-USA-hates agreements will be shredded, increasing the amount of competition.
If it walks like a duck ... (Score:2)
... the claims made here are extremely dubious, and the importance of a home-grown alternative to ASML make this an ideal target for fraudsters.
It is not just ASML which is needed for extreme UV lithography machines. It is a whole ecosystem of European research institutes like IMEC [wikipedia.org] or optics companies like Carl Zeiss [wikipedia.org]. That a startup can credibly match their collective expertise is highly unlikely.