IBM and Google told the Financial Times they now expect to build an industrial-scale quantum computer containing at least one million qubits before the end of the decade. The companies say recent engineering breakthroughs have moved the long-sought goal of practical quantum computing from theory to an attainable target by 2030. IBM in June published a blueprint for scaling its superconducting technology, shifting to a low-density parity-check error-correction code that it says will require about 90 % fewer ancillary qubits than Google’s surface-code approach. The change follows crosstalk interference encountered when the company expanded its ‘Condor’ processor to 433 qubits, underlining the technical hurdles in moving beyond today’s prototypes, which generally contain fewer than 200 qubits. Google, which reported a breakthrough in error-correction late last year, is working to cut component costs tenfold to hit a US$1 billion price point for a full-scale machine. Despite the progress, Amazon Web Services hardware chief Oskar Painter cautioned that commercially useful quantum computers could still be 15 to 30 years away, citing the complexity of packaging millions of fragile qubits and operating them at near-absolute-zero temperatures. The race is drawing new entrants. This week a team led by Pan Jianwei at the University of Science and Technology of China created an array of more than 2,000 rubidium atom qubits, the largest reported for neutral-atom systems. In the UK, start-up Oxford Ionics installed its ‘Quartet’ trapped-ion computer at the National Quantum Computing Centre, adding momentum to a field that is increasingly global even as IBM and Google set the pace.