Trapped-ion quantum pulls ahead on logical qubits; Quantinuum bets on 2030
Helios yields 94 logical qubits from 98 physical; Quantinuum accelerates fault-tolerance to 2030 and signs Mitsubishi; IonQ publishes a 'walking cat' blueprint
Summary
Trapped-ion quantum is winning the logical-qubit race even as superconducting leads on raw count. Quantinuum's Helios system, 98 all-to-all connected physical qubits at 99.921% two-qubit fidelity, yielded 94 logical qubits by March 2026 via efficient error-correction codes, and on 5 June Quantinuum unveiled an accelerated roadmap to universal, fully fault-tolerant computing by 2030, alongside a 2 June MoU with Mitsubishi Electric. Helios couples to Nvidia GB200 over NVQLink/CUDA-Q for real-time error correction, with Amgen, BMW, JPMorgan and SoftBank as launch customers. Rival IonQ published an April fault-tolerance blueprint using "walking cat" codes and reports >99.99% two-qubit fidelity on ytterbium hardware. The gate remains error correction, not qubit headcount, the same logic as IBM's superconducting push.
By the numbers
- 94, logical qubits from Helios's 98 physical qubits (March 2026).
- 99.921%, Helios two-qubit gate fidelity (>99.99% claimed by IonQ).
- 2030, Quantinuum's accelerated universal fault-tolerance target.
- 5 / 2 June 2026, Quantinuum's roadmap announcement / Mitsubishi MoU.
- 4, Helios launch customers (Amgen, BMW, JPMorgan, SoftBank).
Why it matters
Whoever first delivers many high-fidelity logical qubits, not just many physical ones, reaches commercially useful quantum first. Trapped ions' all-to-all connectivity and low error rates make them the efficiency leader; hybrid GPU coupling makes them deployable now. The 2030 target compresses a timeline most of the field still measures in decades.
What to watch
- Independent verification of Helios's 94-logical-qubit claim and scaling beyond.
- IonQ moving its "walking cat" blueprint from paper to hardware.
- Whether trapped-ion clock speed (slower than superconducting) caps throughput.