Oracle Brings Quantum Computing to Cloud Infrastructure With Quantinuum Partnership
AUSTIN, Texas — Oracle Cloud Infrastructure customers will soon be able to run hybrid quantum-classical workloads without leaving their cloud environment. The company announced an 11 August partnership with Quantinuum to deploy the trapped-ion quantum computer Helios inside OCI data centres, making it accessible as a managed cloud service alongside GPUs and high-performance computing clusters.
The deal marks the first time a major hyperscaler has embedded a commercial quantum system directly into its infrastructure rather than offering remote access through a third-party portal. For enterprises exploring quantum advantage, it removes the friction of procuring specialised hardware, negotiating separate contracts, and stitching together disparate governance models. The integration also signals a shift in how cloud providers think about specialised accelerators: not as external services reached over the public internet, but as first-class infrastructure components provisioned, monitored, and billed alongside compute instances.
The Hardware: Helios and the Trapped-Ion Approach
Quantinuum's Helios, launched commercially in November 2025, is a 98-physical-qubit trapped-ion system that has demonstrated 48 logical qubits with an average two-qubit gate fidelity of 99.921 percent. That figure exceeds the widely cited "three 9s" threshold considered a milestone for error-corrected quantum computation. The system uses ytterbium ions trapped in a vacuum chamber and manipulated with precision lasers — a fundamentally different physics from the superconducting circuits that dominate IBM and Google quantum roadmaps.
Unlike superconducting qubits that require dilution refrigerators chilled to near absolute zero, trapped-ion systems operate at room temperature with vacuum chambers and precision laser control. A single Helios unit draws roughly 60 kilowatts — less than one percent of the power budget of leading supercomputers, which can consume 16 to 39 megawatts according to 2026 energy-aware computing research. The ions themselves are naturally identical, eliminating the manufacturing variation that plagues solid-state qubits. Coherence times stretch to seconds rather than microseconds, giving algorithms more room to execute before decoherence scrambles the calculation.
The efficiency gap matters for cloud providers already wrestling with power constraints. Oracle's Mahesh Thiagarajan, executive vice president of OCI, framed the partnership as a way to "improve compute efficiency and energy use" while expanding what customers can solve. The companies plan to preview an OCI quantum service in the coming months, combining Quantinuum's development stack with open-source hybrid-programming frameworks. Quantinuum's TKET compiler and pytket Python library will be supported, alongside emerging standards such as OpenQASM 3 for hybrid workflows.
Why Cloud-Native Quantum Changes the Adoption Curve
Until now, organisations wanting quantum access faced a fragmented ecosystem. IBM, Google, and Rigetti offer cloud queues to superconducting processors. D-Wave provides annealing systems for optimisation. IonQ and Quantinuum have sold access through their own portals or via AWS Braket and Azure Quantum as remote backends. Each path introduces latency, data-transfer overhead, and a separate security boundary. Data must leave the customer's cloud tenancy, traverse the public internet, execute on a remote quantum processor, and return — a round trip that adds seconds to each job and complicates compliance for regulated industries.
Running Helios inside OCI changes that calculus. The quantum processor sits in the same network fabric as the customer's GPU clusters, storage, and identity management. Jobs can pass intermediate results between classical and quantum resources without traversing the public internet. Governance, billing, and access control remain under the customer's existing OCI tenancy. A developer can spin up a quantum endpoint the same way they provision a GPU shape, attach it to their virtual cloud network, and apply the same identity policies.
IDC's Heather West, global quantum research lead, noted that "deploying quantum systems within private cloud environments enables organisations to integrate quantum computing into existing AI and HPC workflows through familiar cloud infrastructure and development tools, reducing barriers to adoption." The Ellison Institute of Technology's Johannes Blaschke echoed the point: having both GPUs and QPUs in one platform "would provide an all-in-one platform, simplify the operation of novel hardware, and help us move at speed from concept to execution."
The Hybrid Workloads Targeting Enterprise Value
The partnership targets use cases where quantum acceleration complements rather than replaces classical compute. Quantinuum and Oracle cited drug discovery, materials science, financial modeling, and large-scale optimisation as initial focus areas. In each domain, the workflow follows a pattern: classical pre-processing prepares the problem, a quantum subroutine tackles a specific sub-task — such as simulating molecular electronic structure or sampling from a complex probability distribution — and classical post-processing integrates the result.
This hybrid loop is already familiar to AI researchers who offload matrix multiplication to GPUs while keeping control logic on CPUs. The quantum variant swaps the accelerator for a QPU. The difference is that quantum subroutines can, for certain problems, explore solution spaces that scale exponentially on classical hardware. Variational quantum eigensolvers for chemistry, quantum approximate optimisation for logistics, and quantum kernel methods for machine learning are among the algorithmic families expected to benefit.
Quantinuum CEO Rajeeb Hazra described the convergence of "quantum, AI, and high-performance computing" as the next phase of enterprise computing. The claim rests on the observation that many scientific and industrial problems — protein folding, catalyst design, portfolio optimisation under complex constraints — already push the limits of today's architectures. A cloud-native quantum layer gives researchers a new lever without requiring them to become quantum hardware experts. The OCI quantum service will expose these capabilities through familiar REST APIs and SDKs, lowering the skill barrier for experimentation.
Power, Supply Chain, and the Infrastructure Reality
The announcement arrives as the data centre industry confronts unprecedented power pressure. Uptime Institute data cited by Cloud Computing News shows that roughly half of the 250 large-scale data centre projects announced globally between 2021 and 2024 face delays or cancellation, with grid access and power availability as primary bottlenecks. The seven largest planned campuses collectively seek 45 gigawatts of onsite power — roughly equal to the UK's entire peak demand. Capacity charges in the PJM Interconnection region have risen more than 1,000 percent between 2024 and 2026, with manufacturers arguing that data centre demand contributes to higher costs.
In this context, a quantum system drawing 60 kilowatts is negligible. But the broader signal is noteworthy: hyperscalers are diversifying their compute portfolio to include specialised accelerators that deliver more answers per watt. Oracle's own cloud business has described itself as "compute-constrained" as AI demand outstrips capacity. Adding quantum capacity that sidesteps the GPU power envelope expands the solution space without proportional grid demand.
Supply chain constraints also favour the trapped-ion approach. Superconducting quantum processors require specialised fabrication foundries with extreme purity and yield requirements — a bottleneck that has limited the pace of qubit-count growth across the industry. Trapped-ion systems rely on vacuum chambers, lasers, and precision optics — components with broader industrial supply chains and more established manufacturing ecosystems. Quantinuum's global footprint across the United States, United Kingdom, Germany, Japan, Qatar, and Singapore further reduces single-point-of-failure risk.
What Comes Next
Oracle plans to preview the OCI quantum service in the coming months. The service will combine Quantinuum's software stack with support for open-source hybrid frameworks, allowing developers to build, test, and refine quantum-classical applications. Pricing, regional availability, and service-level agreements have not been disclosed. Early access programmes for strategic customers in pharmaceuticals, financial services, and government research are expected to begin before general availability.
For the broader cloud market, the partnership sets a precedent. If OCI can operate a quantum computer as a first-class infrastructure service — provisioned, monitored, and billed like a GPU instance — other hyperscalers will face pressure to follow suit. Microsoft already offers Azure Quantum as a marketplace of remote backends. AWS Braket follows a similar model. Google's quantum hardware remains confined to its own labs. Embedding a partner's quantum processor directly into the cloud fabric, with unified governance and networking, is a new tier of integration.
The industry will watch whether enterprise customers adopt the service at scale or treat it as an R&D sandbox. Early adopters in pharmaceuticals, finance, and materials science have the budgets and problem sets to justify experimentation. Universities and research institutes gain access without capital expenditure. The test is whether hybrid quantum-classical workflows become a routine part of the cloud developer toolkit or remain a niche for specialists.
Sources: Oracle Press Release: Quantinuum and Oracle Partner to Accelerate Hybrid Quantum Compute Adoption on Oracle Cloud Infrastructure, Reuters: Oracle, Quantinuum partner to bring quantum computing to cloud
Internal links: Cloud & Edge Computing, AI
Keywords: Oracle, Quantinuum, quantum computing, cloud infrastructure, OCI, Helios, trapped-ion, hybrid quantum-classical, HPC, AI infrastructure