Oracle will put a Quantinuum Helios inside one of its own AI data centers

Quantinuum and Oracle announced a multi-year strategic partnership to bring quantum computing to Oracle Cloud Infrastructure. Helios — Quantinuum's most advanced system, a 98-physical-qubit trapped-ion machine with average two-qubit gate fidelity of 99.921%, past the "three nines" threshold — will be deployed in a US-based OCI AI data center to run hybrid quantum-AI workloads as an OCI service, sitting alongside Oracle's GPU and HPC infrastructure. Oracle plans to preview a managed OCI quantum service in the coming months, with tools for building and running quantum-classical applications, aimed at drug discovery, materials science, financial modelling and large-scale optimization. Quantinuum disclosed the deal alongside its first earnings since its IPO.

Why it matters

Quantum access has mostly meant a machine in a vendor lab reached over an API. Putting the hardware inside a hyperscaler's own AI data center changes the latency and the plumbing between the quantum processor and the GPUs, which is the practical requirement for hybrid workloads where a classical model calls a quantum subroutine inside a loop rather than as a batch job. It is also a distribution deal for Quantinuum at the moment distribution matters most: a newly public company gets its system in front of Oracle's enterprise base without building a cloud, and Oracle gets a differentiator against AWS Braket and Azure Quantum, which broker third-party machines rather than host frontier hardware in their own AI halls. What is not yet established is demand — a preview is not a product, 98 physical qubits is not error-corrected computing, and no customer workload has been named.

What to watch

When the preview opens and at what price, whether any named enterprise workload runs on it, and whether AWS or Microsoft respond by co-locating hardware rather than brokering access.