An investigation into Substrate
substack.com
substack.com
Oof. Before this point in the article I wasn't sure if it's fraud, but comparing the money raised for a (failed) alarm clock vs money raised for supposedly-successful ASML competitor was a good of putting things in perspective. And the nuclear fusion thing... oh god.
US startup Substrate announces chipmaking tool that it says will rival ASML - https://news.ycombinator.com/item?id=45745536 - Oct 2025 (4 comments)
Can Substrate disrupt ASML using particle acceleration? - https://news.ycombinator.com/item?id=45732431 - Oct 2025 (11 comments)
There are some strange things in this writeup-- I don't think it's clear that the machine they claim to intend to sell is supposed to be a direct writing machine, I assumed it was supposed to be like EUV but with X-rays and some kind of special x-ray tolerant photoresist, but the identity of the founders is at least quite damning, the electrostatic chuck thing might be damning I guess, unless there's some special concern. I assumed that even if it were sensible, it wouldn't work well enough, with damage to the resist or something else that manufacturers would find unacceptable, but this isn't my area, so I can't really judge.
“Smart,” I mean, who knows. Sometimes our talents, expertise, and opportunities line up in a way to make us look smart, other times they don’t.
He’s also fairly politically connected and this seems to be a pretty… politically connected project.
IMO the more likely thing is that he’s clever and expecting to get a return-on-investment using expertise that he’s already proven. Not semiconductors, but investor networking.
I’m sticking with my original speculation:
> He’s also fairly politically connected and this seems to be a pretty… politically connected project.
Thiel is, and if he is to realise something political with these guys, then he can only do so if they aren't fraudsters but are actually doing sensible things, or the political project fails. It can of course also be a fakeout for negotiations "we'll have an ASML clone tomorrow, you don't have any leverage" even though they don't, but I don't think so, because if you tried that fakeout on an EU politician, they'll talk to their guys 'is it true?' and those guys will actually ask people at AMSL "is this is a real competitor?"
It doesn’t have to be a centerpiece or anything, so it probably doesn’t have to stand up to much scrutiny from the general public. Like if the government gives them some little grants, they’ll have enough cover to include “revitalizing and onshoring US semiconductors” or whatever in a speech. The press here is not very aggressive about these sorts of claims.
Thiel apparently 'only' has 27 billion USD, so if this is mostly his investment it's 1/27 of all the money he has, and I think when you're spending those sums on highly technical stuff involving physics and precision, you'd have the person representing the project you're planning to invest in, some physicists you know and trust, some people from a relevant industry, and then you sit down and think about the idea as you do when you read a paper, and you'd sit in this seminar for hours and use the blackboard, and my assumption is that you'd go away very close to knowing for sure whether it would work or not.
Maybe reconsider: https://www.theguardian.com/us-news/2025/oct/10/peter-thiel-...
They tend to be crazy ideas with fundamental flaws and either don't happen are are silently reshaped to be feasible.
The "clever" interpretation would be that this is primarily done to fleece the states treasury, and I'm certain that still happens, but I firmly believe that the driving force is one ultra rich ultra powerful naive person without any negative feedback from his inner circle.
If they draw something on a piece of paper, believing themselves to be very visionary and order that thing to be built people will just start doing so without a word of warning. Speak up and you might find yourself short of a few limbs.
Trying to trace things like this is a fools errand though. Trace things back further than that, the microscope was invented in Europe, and further back the magnifying glass somewhere in the Middle East/Mediterranean. You'll just stop wherever it's convenient for the point you're trying to make.
Ultimately, what was available + what was known was put into practice by ASML and they're the ones best at executing that thing right now today, and some things they're the only ones able to too, which isn't very too common.
But I guess only time will tell if they can use whatever they've built and make it into something even more.
It is not. EUV was literally invented in America in government funded research, and licensed to ASML. Taiwan (TSMC) purchased ASML machines.
> For all of America's DARPA prowess, Intel couldn't even play second-fiddle to TSMC's worst fab nodes.
These have nothing to do with each other. One is a private corporation that made a strategic mistake. The other is a government agency.
Modern, high-yield EULV was not. The CRADA efforts did not produce high-yield dies, nor did it get beyond theorizing modern nodes.
> These have nothing to do with each other
Maybe that's how you wish things were. But we have two admins back-to-back scraping Intel's corpse off the pavement.
If you have a source that says "modern, high-yield EULV" does not use ASML machines, or does not use the laser pulsed tin plasma light source invented in the US, I'd be interested
Claiming that America invented modern high-yield EUV in the lab is like saying Germany invented the Ferrari because Italy imported their steel presses to manufacture it. Not only are you being pedantic, you're not using pedantry to make a distinction that anyone will respect. Chipmaking is not a pushbutton process with line-replaceable units. This kind of reductive reasoning is what puts detail-oriented cultures ahead of profit-oriented ones.
Regardless of what other prior inventions it depended on, EUV technology itself was invented in the US, and we know that factually - it’s not a fool’s errand. ASML literally approached the US government for a chance to license the technology. This was in the late 90s. That’s not to take away the work ASML has done since then to make it commercially viable through their machines (or TSMC in implementing and integrating ASML machines).
Also, process fabs are volume driven technologies, and focus on getting as much product out the door as quickly as possible. A lot of technologies will never be compatible with that optimization in a business context.
Samsung was sort of the exception to this trend, and took huge risks that paid off =3
https://newsletter.semianalysis.com/p/how-to-kill-2-monopoli...
Skeptics welcome to review the following section:
> These are extraordinary claims and thus demand extraordinary evidence. Let’s take them one-by-one:
And here is their conclusion:
> 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. Still, they have at least developed some impressive capabilities on the most complex part of the process (litho) in a short amount of time (2-3 years).
It's easy to get stuck judging people based on their history, so you easily skip over other details that might make their current efforts different, we all think (hope?) we grow, meaning so does this founder, despite past failures. I'm not trying to say that this is a signal either way, but I wish others don't judge me in the same basic way in the future, as I too have failed in the past (although not about solving nuclear fusion) but think I'm getting better every day. Of course, people will judge me and others either way, but it seems more reasonable to not not just jump into history first, and focus on other details that actually evaluate the ideas themselves.
Can Substrate disrupt ASML using particle acceleration?
https://www.nytimes.com/2025/10/28/technology/can-a-start-up... (https://news.ycombinator.com/item?id=45732431)
See also:
Return to Silicon Valley
Haha, am I still supposed to believe VC’s are competent, intelligent people who deserve their stature when they fall for scams that should be more obvious than a Nigerian prince scam.
> But yes, you’d expect more due diligence from investors, but a lot of times it’s just about how likeable you are and who you know.
I think both of these can be true simultaneously. The limits of due diligence and the practical ability to evaluate an investment opportunity (quickly and at low cost) mean you sometimes do make mistakes, even “stupid” ones. But also, access and networks and old boys clubs are rampant in the VC ecosystem.
It’s as fair as characterizing an athlete saying he’s competing to win.
A common logical/semantic mistake is that claims about future facts are lies. They can be exaggerations, and we do see this often and I’m personally annoyed when I see it. These claims can be genuine or not. I often think it’s easier for a founder to be naive (or crazy) than not, because they can believe (sometimes strongly) the things they say.
But neither are lies. A lie is when you misrepresent a fact, something that already happened. That’s a big causal difference.
The large gray area is when someone misrepresents intent (as with most cases for cons). This, I suspect, is the main questions you should be asking, but I suspect it falls into the ethics rather than legal debates.
There absolutely is a promise. Even if you manage to legally find a way to not get sued, taking advantage of the fact that everyone who gave you money believed it was a promise is still scamming them.
* Promised an alarm clock that would do a bunch of things
* Took $2.5M in funding from Kickstarter
* Took another $50M in funding from elsewhere
* Delivered a piece of hardware that did essentially none of what was promised.
It's all detailed in OP and the linked Verge article. That's a scam and I'm not interested in your legalese arguing whether they can be sued or not.
What... You know, it doesn't matter. Thanks for the summary anyways!