John F. Clauser is a celebrity now that he’s been awarded the Nobel Prize in Physics, but when Louisa Gilder interviewed him for her book The Age of Entanglement : When Quantum Physics Was Reborn, he was just another ragged experimentalist toiling away in his lab.
There was something he said in that interview which stuck in my mind, here’s what he said:
To be an experimental physicist, you need to be able to make anything. You need a mill and a lathe. But most of all, you need a junkyard. The most valuable commodity in any physics department is floor space.
This resonates with me, since I’ve lately spent a lot of time with experimental physicists and have come to appreciate just how dependent they really are on having space.
The problem of course, is that good cleanrooms and nano-fabrication labs are prohibitively expensive to build and keep up to date.
That means access to such facilities is always going to be a bottleneck. Especially so, if you grow out of the university context—where lab access may be scarce but at least it’s free—and want to try doing the entrepreneurial thing.
One option of course, is to go “fabless”, meaning that you stick exclusively to designing devices and let some other entity do the fabrication.
This model has many advantages. “Designed in California. Assembled in China” has worked wonders for Apple. Nvidia is fabless too.
It’s difficult to start fabless thought, since innovation requires a lot of tinkering to see what works.
That’s why granting generous access to university lab facilities is such a powerful lever, if the ambition is to encourage entrepreneurship in strategically important fields. Fields such as microchip design or quantum technology. Fields which promise to create immense value for the innovation clusters that create the best conditions for enabling innovation.
Treating lab access as infrastructure, a public good, seems to be a key factor in such enabling.
And sure enough, there’s been a number of such initiatives. In the US, we’ve got the NACK Network Nano4me and the nanoHUB RAIN lab. In Sweden there is MyFab.
While those are all commendable efforts, none of them really seems to fit the bill.
That impression was anecdotally validated earlier this week, when I had a chat with the founders of two of Sweden’s fastest growing quantum tech startups.
Neither of these companies would have come into existence if it wasn’t for subsidised access to university lab facilities, but both are now looking at investing in their own proprietary fabrication labs. They feel they have to because, as they put it: “A shared facility like MyFab is great for the very early stages, but it’s not really built for supporting scale-ups”.
Let’s say these two companies are successful in building up their own machine parks, then where are we at? I came across a piece of news the other day that Sweden’s only proper MEMS foundry—Silex Microsystem—was sold to a Chinese company with close ties to the government. Laws have since been passed which would have prevented such an acquisition, but that doesn’t help in retrospect and laws can also be circumvented. As long as unique infrastructure belong to privately held companies, it’s going to be vulnerable for takeover.
I think it would be so much smarter if we could incentivise up-and-coming players to share their physical infrastructure. To compete on ideas and execution, rather than on who’s got the bigger machine.
I wonder what it would take for a national, or even cross-national network of cleanrooms and fabrication facilities, to be so good and accessible, that startups don’t feel they need to invest on their own.
I think figuring out the answer to that question is probably going to be pretty important.