Only 10,000 qubits to break public key cryptography: researchers' predictions

A few years ago, it was believed that millions of physical qubits were needed for quantum computers to perform useful calculations for scientific research and industrial applications, as well as to break the public key ciphers used today. However, according to research from Caltech and Oratomic, a startup linked to the Californian university, it is possible that far fewer qubits are actually needed: it may be feasible to build a "useful" quantum computer with just 10,000 or 20,000 qubits.

Are "useful" quantum computers arriving sooner than expected?

The basic problem that currently available quantum computers face is that of errors: the delicate quantum processes underlying the operations of these computers are extremely susceptible to interactions with the external environment, which cause them to end with errors. This is why we speak of "physical qubits" and "logical qubits": the former are the actual qubits present in the computer, subject to errors and interference, while the latter are decoupled from the physical component and are the qubits seen by programmers, without errors.

Building logical qubits, however, requires multiple physical qubits. The general idea is similar to that of RAID in the classical world: multiple qubits store the same value; if one encounters an error, it can be corrected by comparing it to the others. Researchers have continued to reduce the number of physical qubits needed over the years, from 10,000 to 2,000, down to a few hundred, and even down to just 5 in Caltech's research.

According to the study, it is possible to use neutral (or Rydberg) atoms to achieve this. The key, as explained in the study published on arXiv, lies in the ability of neutral atom-based computers to move atoms at will using "optical tweezers" (extremely precise lasers): this allows for the distribution and entanglement of distant atoms, thereby reducing the number of atoms needed to achieve error protection.

The result obtained by the researchers is purely theoretical for now, but these are the same scientists who built the world's quantum computer with the most qubits, reaching 6,100, and they believe that devices with 10,000 qubits and more are expected by the end of the decade.

Predictions from Google and Cloudflare indicate that we should prepare for the breaking of public key ciphers by 2029. It remains uncertain whether this will definitely occur, and whether (if at all) it will be due to neutral atom computers, but the entire sector is moving rapidly towards reducing the number of physical qubits required to create a logical one, and therefore also the number necessary for practical large-scale applications (given that scientific discoveries have already been made using quantum computers).

The need for companies to take action to implement post-quantum cryptography is becoming increasingly urgent. Companies will need to act not only to update their software but also to mobilize their entire supply chain. We have quickly moved from a purely hypothetical situation far into the future to one that could be just around the corner, relatively speaking.

It's time to lay out clear and concrete plans.