The competition to build a useful quantum computer is heating up. In September, Microsoft and Atom Computing announced that they set a record for the largest number of entangled logical qubits. This result, which points towards a quantum computer’s ability to correct its own errors, has now been bested by the start-up firm Quantinuum. The company says it has made the most logical qubits yet.
All computers make errors, and catching and correcting them is crucial for their operation. But this task is more difficult for quantum computers than for conventional ones: the laws of quantum physics make it impossible for qubits – the quantum equivalents of bits – to mimic conventional error-correction procedures. Instead, researchers group qubits into “logical qubits”, which can then be used to run programs with much lower error rates.
To use logical qubits, researchers often have to entangle them first. Atom Computing and Microsoft previously used a quantum computer made from extremely cold ytterbium atoms to create and entangle 24 logical qubits. Quantinuum researchers have more than doubled that number.
Advertisement
They created and entangled 50 logical qubits made from cooled and electrically charged ytterbium atoms, according to at Quantinuum. He presented the work on 10 December at the Q2B conference in California.
The result breaks two records: it is the largest number of logical qubits – a title previously held by researchers at Harvard University and the quantum start-up QuEra for the 48 they created in 2023 – and the largest number of entangled ones.
Hayes says that even though this set is the biggest yet, the specific quantum state that they encode can be used to detect errors a quantum computer makes during calculations, but not necessarily to correct them. In this way, the record-breaking state is an intermediate step on the road towards fault-tolerant quantum computing where all errors would be corrected.
Researchers currently disagree on which quantum computer design will be able to reach fault tolerance, but whichever does will be the first to deliver on the promise of quantum computers outperforming the world’s best supercomputers. While there is broad agreement that logical qubits will play a big role in getting there, various research teams have taken different approaches for building and using them too.
For instance, earlier this year researchers at Google Quantum AI demonstrated a method for making a single logical qubit become more reliable with increasing size using a quantum computer with tiny superconducting circuits as qubits. This marks two competing approaches to error correction – making more and more less-than-perfect logical qubits, or perfecting one larger and larger logical qubit. With records being broken so often, it is still unclear which approach will win out.
