The Signal
IQM Entangles All 54 Qubits on Its Emerald Processor
IQM says its cloud-accessible Emerald quantum computer prepared a Greenberger–Horne–Zeilinger state across all 54 qubits at once. The company reports a fidelity of 0.516 with readout error mitigation and says cloud customers can use the ful
Quantum.world Team · Written with AI · fact-checked ·
Emerald’s full chip enters one entangled state
IQM Quantum Computers said on October 29 that it had entangled all 54 qubits on its cloud-accessible Emerald processor at the same time. The company used a Greenberger–Horne–Zeilinger (GHZ) state, a standard benchmark in which quantum correlations extend across the full group of qubits.
IQM reported a fidelity of 0.516 for the 54-qubit result, above the 0.500 threshold it identifies as the boundary for genuine entanglement. That full-chip figure includes readout error mitigation (REM), a correction applied to measurement results. The company also reported a fidelity of 0.515 at 40 qubits without REM, so the two figures do not provide a like-for-like comparison of performance with and without that correction.
The announcement marks a change from what IQM says users could previously achieve on Emerald: around 30 qubits in a comparable demonstration within one algorithm. The company says cloud customers can now access the full-chip performance. A GHZ-state demonstration establishes entanglement across the processor; by itself, it does not show that the system outperforms classical computers on a useful application.
An automated routine chooses gates and timing
IQM attributes the result to an automated state-preparation routine that selects qubits and two-qubit gates while balancing gate fidelity against idle time. The order of operations matters because qubits can lose useful quantum information while waiting for other gates to run. The routine aims to reduce that idle time as it builds the GHZ state.
“Entangling the entire chip means every qubit and every connection between them has to work correctly in the same circuit. Getting there came down to being smarter about how we prepare the state itself, choosing the gate order to keep idle time to a minimum,” said Craig Ciesla, CTO of IQM Quantum Computers.
The company says the work also draws on about a year of hardware and calibration changes, including an updated calibration methodology, an improved gate set, dynamical decoupling and qubit selection. IQM says those improvements reduced error rates for quantum operations on the processor by up to 50%. The release does not specify the baseline or measurement conditions behind that maximum reduction.
The fidelity figure needs its measurement context
IQM describes the result as among the strongest reported for a commercially available quantum computer, but its announcement does not provide a named comparison set or a head-to-head benchmark. The reported 0.516 figure is therefore best read in the context the company gives: a 54-qubit GHZ-state measurement with REM, above the stated entanglement threshold.
That distinction matters because a processor-wide entanglement test answers a narrower question than a general-purpose computing benchmark. It shows that IQM’s calibration and state-preparation methods can bring all 54 qubits into a single entangled state under the reported conditions; it does not establish the fidelity of every circuit or workload users might run.
Cloud users get access to the 54-qubit run
IQM says cloud customers can use the same full-chip performance demonstrated in the result. Its release places the automated routine alongside the company’s work on AI-driven agentic calibration, part of a broader effort to automate quantum-system tuning.
For Emerald, the headline number is 54 qubits operating together in a GHZ-state demonstration, with reported fidelity of 0.516 after readout error mitigation. IQM’s release identifies that measurement condition explicitly; the reported 40-qubit result, by contrast, is 0.515 without REM.