Quantum computers have spent years as exotic guests in supercomputing centers. They sat in the building but outside the workflow, wired in with custom integrations built site by site. Three announcements on November 16 chip away at that, from job scheduling to control electronics to error-corrected qubits.
IQM and NVIDIA put QPUs in the Slurm queue
Slurm is the open-source workload manager most big supercomputers use to hand out CPUs and GPUs. Quantum processors never fit its scheduling models. NVIDIA, which now backs Slurm and has committed to keeping it open source and vendor neutral, is changing that. Slurm 26.11 adds plugin support for Hierarchical Resources (HRES), so a center can treat a QPU as a shared, cluster-level resource and hook it up to outside services such as the device’s own control software.
IQM said at SC26 that it supports the work. Its systems already run at Oak Ridge National Laboratory, VTT in Finland, LRZ in Germany and CINECA in Italy, where researchers submit hybrid quantum-classical jobs through the same queues they use for classical work. IQM partner MQSC (Munich Quantum Software Company) has put an open-source MQSC Orchestrator on GitHub that combines HRES with the Quantum Device Management Interface (QDMI) to schedule IQM machines such as IQM Emerald. There’s no general availability date yet. IQM will validate the integration with centers in staging environments as they upgrade Slurm.
“Quantum computers become useful at scale when they behave like every other resource in a supercomputing center,” said Craig Ciesla, IQM’s CTO.
Altera, OQC and NVIDIA build a controller reference design
Altera showed an FPGA-based quantum system controller reference design built with OQC and NVIDIA. It runs on the Agilex 9 Direct RF FPGA, which has the analog-to-digital and digital-to-analog converters on chip, so qubit control and readout skip external RF up- and down-conversion. Frequency multiplexing lets several readout channels share one RF port, which cuts RF hardware as qubit counts grow.
The split of work is tidy. The FPGA handles real-time qubit control, resonator readout and spectroscopy. OQC supplies the RF conditioning electronics between the FPGA and its new-generation QPU. NVIDIA GPUs, linked through NVQLink, do the heavy lifting such as calibration and a machine-learning readout classifier, with CUDA functions driving the FPGA logic. An emulation path lets developers bring up the controller before they have any qubits. The design is due in January 2027.
Infleqtion reaches 30 logical qubits
Infleqtion ran 30 entangled logical qubits in a single computation on its Sqale neutral-atom system, using 80 physical qubits. An AI-assisted discovery halved the physical gates needed for one key logical operation, and fewer gates means fewer chances for error. The company is aiming for 100 logical qubits in 2028.
The business side is still small. Q3 revenue was $9.1 million, up 25% year on year, and nine-month revenue reached $32.6 million, up 57%. The quarter swung to a gross loss of $1.1 million, with a net loss of $19.7 million. Infleqtion kept its 2026 revenue outlook of at least $45.1 million and ended the quarter with $560 million in cash and securities and no debt. A $20 million NASA contract expansion brings its quantum gravity sensing work with JPL to $40 million in total.
These are plumbing stories, really. Scheduling, control hardware and error correction are the dull parts that decide whether quantum machines turn into tools HPC centers actually use. For the algorithm side, see our piece on OQFT and the optimistic quantum Fourier transform.