You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
“Substrate” has been in the news for, allegedly, coming up with a way to make computer chips for much cheaper and at much higher quality than anyone else in the industry.
There’s just one problem: all signs indicate that the entire business is fake.
- The founder is known con-artist involved in such other things as solving nuclear fusion and stealing $2.5M in a Kickstarter scam.
- The cofounder is the founder’s brother and has literally zero documented professional or academic experience.
- The company’s job postings are nonsensical and AI-generated
- The company is unwilling to evidence any of its (extremely extraordinary) claims, which have been made on a timescale that makes no sense in the context of the semiconductor industry
- The company’s research facility is, based on photographs that they’ve published, at least two orders of magnitude smaller than what would be necessary.
Ebeam lithography and multi-beam has been in existence for years (maybe decades) but they are used to make masks not chips.
If the requirement is to make one chip per week per machine, yeah, certainly possible (may even improves to 2 chips a week, lol). But you better be prepared to build fabs the size of superdome or bigger.
Or unless he might have figured out how to flatten a chip design to just one single layer, without the need of muiltiple layers needing mask and exposure.
It's hard to tell -- there is entanglement though of some kind, per Semianalysis:
" Note we have worked with Substrate since as far back as 2022, but the technical analysis here was by team members who did not have access to that NDA information."
Substrate X-Ray Lithography, a New American Foundry, $10k Logic Wafers
newsletter.semianalysis.com
... Further down, this article does point out challenges:
Naysayers will point out a million reasons why this is improbable, difficult, etc. - and they are mostly correct. There is a big difference between lab-scale and industrialized, high-volume tools. Substrate itself realizes this and agrees they are in for a lot of development and scaling pain.
... and their summary doesn't read like they're part of this:
Most of this report is based on the promise of Substrate. A novel demo tool with these imaging capabilities built in 2 years is impressive. But there is a lot more to prove before they disrupt the chipmaking industry. We're hopeful for success but skeptical given how many questions there are. We’re looking forward to a few milestones that will silence the skeptics (both internal and external):
substack.com
