David Chisnall (*Now with 50% more sarcasm!*) on Nostr: The economics of quantum computers for breaking encryption are interesting. Although ...

The economics of quantum computers for breaking encryption are interesting.

Although the most exciting thing you can do with a quantum computer (of the approximately two things people have figured out how to do with them so far[1]) is simulating quantum chemistry. But that's not where quantum computing funding comes from.

The high-profile use case is breaking encryption. And the market for that is:<li>Spooks.</li><li>Law enforcement.</li><li>Criminals.</li>

PQC adoption changes their interest level significantly. Spooks are mostly concerned with nation-state and state-sponsored folks and with terrorists. Most of the first categories are already moving to PQC. A realistic quantum computer in three or four years will mostly give them five-or-more year old data. Useful? Yes. Groundbreaking? Not so much. Adoption of PQC likely removes them as a funding source.

Law enforcement buys these things, but doesn't invest in the R&D. That money may come from companies that aim to sell to law enforcement. But law enforcement rarely cares about very old data, so adopting PQC in consumer devices kills that as a route to market.

Criminals will always go after the very long tail of late adopters. They're one of the major funders of Meta (10% of revenue came from ads for scams, per the recent report), for example, and big customers of AI tools. So they may be customers of this kind of thing, but probably not because quantum computers are more expensive than the many other things that work. And creating a business where your only customers are criminals is fragile.

Which leads to an interesting situation that, if the world adopts PQC rapidly, then that eliminates a lot of the financial incentives to develop quantum computers. And that means adopting PQC looks like a waste of money.

[1] Quantum AI is a buzzword to con investors, not a technology. Quantum-annealing machines are already real and are useful for a bunch of simulators but aren't 'true' quantum computers.