29 stories
Show: Summaries / Headlines only
IBM and University of Chicago researchers claimed advantage with a certified fidelity bound on 97 physical qubits. Qedma and Algorithmiq reported regimes in which the classical methods they tested became unreliable or disagreed. None of the preprints is peer reviewed.
McKinsey built the range from expert interviews and 2035 baselines for seven finance segments. It footnoted the use cases' impact levers as the impact of quantum computing and AI. McKinsey also called its value estimates approximate.
The team sent the largest fragments of small molten-salt clusters to an IBM processor, using circuits of up to 66 qubits plus ancilla qubits. The authors traced the dominant remaining error to the classical step that split the clusters.
Buyers already pair control electronics and refrigerators from different suppliers and use open compiler formats. Calibration has a published specification only for pulse jobs. Module links have none. In those two layers, buyers still depend on one vendor.
A quantum computer runs on clocks from nanoseconds to hours. IBM and Xanadu both drew their software layers along those clocks. A buyer still cannot swap one vendor's processor, controller or decoder for another's by recompiling.
IBM listed five spending areas, from research to acquisitions, without assigning amounts. Commerce had signed a letter of intent for $1 billion in planned CHIPS funding toward Anderon, IBM's planned quantum wafer foundry.
Teams publishing in Science measured quantum learning advantages on quantum data, including an 11.8-order cut in samples. On ordinary datasets, classical models led on four of five in a peer-reviewed benchmark. HSBC and IBM reported gains in a preprint.
Silicon spin qubits now come from 300 mm lines at imec, Intel, and STMicroelectronics. No maker has yet shown a fab supplying enriched silicon, qubits uniform enough to tune automatically at scale, or control electronics within a million-qubit power budget.
Thirteen IBM researchers laid out how quantum processors should share schedulers, networks, and security controls with CPUs and GPUs. They described early installations at RIKEN and RPI. Tighter coupling and fully co-designed systems remain proposals.
