Nord Quantique's Smartqubit
(They're not calling it that.)
Roughly 2.5 decades ago (that hurts to write) I double-wielded a Motorola StarTAC flip phone with an Internet-capable Palm VII PDA (personal digital assistant). No, AI doesn’t render them great, but work with me here, people. My next phone combined their capabilities into an early model smartphone, allowing both calls and apps with one device. Over the decades (ouch) since, this form factor has become significantly more powerful and feature rich.
And that’s my impression of Nord Quantique’s multimode bosonic qubits in a nutshell. A walnut shell, to be precise.
Instead of scaling up like everyone else, adding lots of qubits to implement quantum error correction codes, Nord Quantique is essentially scaling down. Keeping the same form factor, like a smartphone, these walnut-sized qubits are further developing on the inside. Therefore, as logical error rates come down, the size of the system won’t change.
Peer Review: Check.
Nord Quantique popped up on my radar at the same time it popped up on one of Prof. John Preskill’s slides. I personally consider that the ultimate peer-review, but if you’re a stickler for reading papers, then obviously we know that Prof. Preskill read peer-reviewed material. There is also a whitepaper, which is what I had read before asking this question. There is also some new material that has not been peer-reviewed yet, but earlier material has.
Hype Detector: On.
Let’s check out some of Nord Quantique’s claims:
Lower requirements for cryogenics?
The qubits need to be as cold as any other superconducting qubits, but the size of the system will be smaller. Everything will be able to fit into one fridge, with no interconnected systems, so the total cryogenic demand will be lower than other superconducting systems.
Smaller physical systems?
Depending on the size of the fridge, today’s tech will allow more than 2,000 logical qubits in just one fridge. That’s 2,000 logical qubits, ladies and gentlemen, occupying only about 20 square meters, which is compact enough to integrate inside a datacenter.
Energy savings?
The main energy driving factor in quantum computing is the classical control electronics. Besides the aforementioned reduction in cryogenics, less control is needed. Whereas multimode bosonic qubits are more complicated than transmons, they’re much less complicated than controlling thousands of transmons.
Why RSA-830?
I thought we cared about RSA-2048, so I was curious why RSA-830 was chosen for a comparison. The answer is that RSA-830 can be compared to a supercomputer, whereas RSA-2048 cannot. Nord Quantique’s estimate is that it will be able to break RSA-830 in one hour at a speed of 1 MHz, consuming 120 kWh. The HPC estimate is 9 days at a rate of 1,300 kWh, consuming 280,000 kWh.
Conclusion
Nord Quantique’s first utility-scale quantum computers, with more than a hundred logical qubits, are promised by 2029. Thanks to smartphone-style “scaling down,” as I’m calling it, that’ll be 100 logical qubits in just one fridge: a smaller and greener physical footprint for superconducting quantum computers.
Image generated by an AI model provided by Microsoft Copilot.