QuEra has significantly reduced its qubit error rate—its first commercially available computer will launch this year (with 256 physical qubits and 10 logical qubits).
Logical qubits (qubits connected through quantum entanglement) reduce errors in quantum computers by storing the same data in different places. This diversifies the failure points when performing calculations.
The new QuEra machine, with 256 physical and 10 logical qubits, will be launched in late 2024, an announcement that follows the publication in the journal Nature of a paper by researchers from Harvard, QuEra and several other institutions that demonstrated a functioning quantum computer , which contained 48 logical qubits (the largest number of logical qubits tested to date).
“This is the first quantum error correction machine,” says study co-author Harry Zhou, a physicist at QuEra and Harvard University.
While the computer doesn't have enough power to be useful on its own, it does provide a platform on which programmers can start testing code for future quantum computers, Zhou said.
Why Quantum Computing Requires Error Correction?
While conventional computers store information in bits with a value of 0 or 1, quantum computers use qubits – which are a superposition between 0 and 1, using the laws of quantum mechanics.
Qubits can also be stitched together using quantum entanglement and exist in multiple states at the same time—this allows them to perform many calculations much faster than classical computers. But qubits can be easily broken, making them error-prone: about 1 in 1,000 qubits fail, versus 1 in 1 billion bits in conventional computers.
Art Director course. Gamedev. In order to upgrade your skills and become an art director in a game device. More about the course
The largest quantum computer built to date has only about 1000 qubits, and its high failure rate limits its potential expansion. Error correction can counteract the tendency for qubits to fail, and creating logic qubits is one of the most effective methods.
Quantum noise reduction
According to Zhou, the new error correction system is based on data redundancy, where the same piece of data is stored in multiple places. That is, logical qubits perform the same calculations on multiple physical qubits, greatly reducing the error rate—if one or more physical qubits fail, the calculations can continue because the data is available elsewhere.
To create a logic qubit, researchers apply error-correcting computer code to ordinary qubits; then install logic gates, or circuits, between the qubits to entangle them. Next, the quantum computer calculates the “syndrome” – whether an error occurred or not, and using this information, it corrects the errors and moves on to the next step.
The new qubits show significant progress over previous efforts. In 2023, Google Quantum AI Lab showed an error rate of 2.9% using three logic qubits – while Quera's error rate was 0.5% with 48 logic qubits. Oxford University is a world leader in this – with an error rate of less than 0.91% (but only between two-qubit gates).
Last year, IBM also demonstrated its own technology in the 127-qubit Heron chip, which reduced the number of errors by 5 times compared to other chips. However, the company's first fail-safe machine for commercial use will not appear until 2029.
QuEra, meanwhile, plans to launch several quantum computers in the coming years, starting with a machine with 30 logical and 3000 physical qubits in 2025. A more powerful computer with 10,000 physical and 100 logical qubits is planned for 2026.
“A machine with 100 logical qubits can perform correct calculations that exceed the capabilities of modern supercomputers,” Zhou said.