General
Looking into substrate - maybe a fradulent business with a questionalble leadership
The startup backed by some people who never have produced semiconductors, know the industry demands deeply and are obviously lacking scientific knowledge is prevalent when you read about it and inform yourself about what they try to do and how they try to approach it, look into some interviews. Some interesting talk about it...
https://news.ycombinator.com/item?id=45767013&utm_source=chatgpt.com
https://substrate.com/our-purpose
To me it looks like some people with money don't know more efficient ways to burn their money. Seems fraudulent. It's much different that a "software business" if you look into production and have seen a clean room from the inside. I think they will fail heavily from the info I read and things I know about it. To get the boots of the ground the money invested is peanuts. The picture with some famous physicists from the past on the website looks like a bad joke for me in this regard. Substantial things are missing besides marketing language and we have something that works. I don't see them building a functioning machine producing high yield chips constantly in the future, catching up with other company's. I question hard that they will understand the complexity of executing high yield production and precession engineering needed.
For the more technical interested:
https://newsletter.semianalysis.com/p/how-to-kill-2-monopolies-with-1-tool
Comment on article:
What's somehow interesting is the framing of where the EUV technology got developed and the (nationalist) thinking of the people writing the article describing complexity with the process. They are wrong from my research and the interconnections of company's in the field, if you trace back some things and look into patents, how ASML build EUV machines and even bought parts of Zeiss to push them when they said they can't technically build parts of the machine for them to get them further money to do it. It's not something one country or state agency has developed. I think it's largely politically/ideology driven and not by scientists, people who can engineer and do the physics. The history and development of EUV lithography is gloriously messy, multinational, and fundamentally engineering-driven, not nationalist hero-myth bullshit stuff. The only consistent narrative that fits the data is EUV exists because thousands of specialists across multiple countries collaborated through a ridiculous sequence of improbable small technical wins and decade long collaboration.
Added form comment bellow:
The problem is substrate misses all of the economics needed, ignores the core fundamental physics problems associated with using x-rays(lenses, continuous and precise high energy sources for continuing production/ the energy costs, particle scattering, creation of radicals, the scale of particle accelerators for creating the x-ray needed, the continuous energy density.) and many more principles of manufacturing semiconductors at this small scales and basic safety problems with using that energy x-ray density on wafers, blur effects and much much more. (Some of this are the reason people decided to don't go that way in the past and do EUV ) What they did is like oh we build a car and it drives exactly one second and then never functions again. It's not thought through from a first principle basis, can't be build like a crappy software startup. Besides that all the big chip producers won't trust them because they have no experience, aren't reliable, when you look into people of the company and their background. I would not bet 20 Billion dollars on them as chip maker. My take is they just funnel money from the chips act/government/ politicians people who don't know the physics. Compact, cheap x-ray lithography with that investment won't be possible and feasible. Anything that cannot sustain thousands of hours of continuous operation at <1 nm vibration noise is dead on arrival for this. The startup model (iterate fast, break things) fundamentally misfits with lithography, where the real model is: Spend 25-30 years with Zeiss and contractors polishing mirrors and tech to atomic perfection and burn billions proving each subsystem before even lunching main product. Taking over company's and forcing them to further develop tech all while continually burning money. They won't solve that as a startup. It shows a lack of understanding for me for a lot of basic things and knowledge. Besides all that the cost per wafer curve just explodes with that technique.
Substrates problems:
zero track record
non-existent supply chain for tech
no experience with building complex systems
no optics heritage and experience with x-rays
no metrology competence
no reliability data
no uptime guarantees
no 24/7 field service network
no 15-year service roadmap
My conclusion: x-ray lithography for high-volume semiconductor manufacturing is substantially more difficult than EUV lithography on nearly every physically relevant axis for implementation (optics, sources, statistics, materials, system integration, and economics). I won't bet on "a startup" doing it.