What’s happening
Quantum computer maker Quantinuum proved the merits of 1990s idea, catching and correcting errors as a calculation progresses.
Why it matters
This key step ultimately could potentially make the revolutionary machines practical, tackling problems in areas like materials science and medicine that are out of the reach of conventional computers.
Quantinuum, a leader in the nascent field of quantum computing, said Thursday it advanced a key technique for correcting errors in calculations done by the advanced machines, a development essential for them to fulfill their revolutionary potential.
A Broomfield, Colorado-based team improved its handling of qubits, the machines’ fundamental data storage and processing elements. Last year, they linked multiple ordinary qubits into a group called a logical qubit that’s more reliable. This year, they got a pair of logical qubits to perform calculations, said David Hayes, leader of the company’s theory group.
The work was notable because Quantinuum’s error correction technology keeps logical qubits stable longer than conventional qubits, and that’s key to coaxing useful work out of quantum computers.
The new achievement could eventually lead to developments that allow quantum computers to crack encryption, design more efficient solar panels and meet other promises of the complex machines. Such potential will likely encourage companies, governments and universities to continue investing in the technology though realizing its potential remains years away.
Quantinuum, created when Cambridge Quantum and Honeywell Quantum Solutions merged in 2021, isn’t the only company working on the idea. Big companies like Google, IBM, Intel and Microsoft are racing with startups like Rigetti Computing, IonQ and others to build the first practical, powerful quantum computers.
Qubits are individual atoms or other elements tiny enough to be dominated by the weird rules of quantum physics. It’s those rules that could help quantum computers solve problems that are beyond classical computers. On Quantinuum’s H1 quantum computers, that means prodding up to 20 electrically charged ytterbium atoms around with a laser inside a cold, airless chamber — an “ion trap.”
One problem with quantum computers is that qubits are hard to control and easily perturbed, derailing calculations. Logical qubits offer a mechanism to stabilize the situation.
Some qubits in the logical qubit flock are used for the data processing while others are used to interact with an ordinary outside computer. That conventional machine oversees the operation by monitoring the logical qubit’s state, checking for errors, and shepherding any errant qubits back into the fold. Quantinuum’s system repeats those error correction steps about five times a second.
With its results, Quantinuum has shown error correction ideas developed in the 1990s really are practical, said Russell Stutz, leader of Quantinuum’s commercialization efforts.
“There’s a path forward to getting lower error rates,” Stutz said.
A calculation with two logical qubits is still an early stage toward powerful, fault-tolerant quantum computers. Quantinuum’s test performed a simple test chore.
“You need more logical qubits to get to the level where this computer can do something that your classical computer can’t simulate,” Hayes said. He estimates about 50 are needed, and that, in turn, means hundreds of underlying physical qubits. That’s beyond Quantinuum today, though quantum computer makers are gradually headed that direction.
