Oracle and Quantinuum Partner to Wire Quantum Computers Directly Into AI Data Centers

A machine that draws roughly 60 kilowatts of power—less than the HVAC system of a mid-sized office building—is about to be physically wired into one of the world's most power-hungry environments: a U.S.-based Oracle artificial intelligence data center.[4]

The integration is the centerpiece of a multi-year partnership announced Tuesday between cloud giant Oracle and quantum hardware maker Quantinuum.[1][3]

Under the agreement, Oracle Cloud Infrastructure (OCI) will host Quantinuum's third-generation Helios quantum computer on-site.[1][5]

The goal is to allow enterprise developers to seamlessly route computational traffic between Oracle's massive clusters of classical graphics processing units (GPUs) and Quantinuum's quantum processing unit (QPU).[1][2]

While the technology industry has spent the last two years fixated on scaling up classical AI models, the Oracle-Quantinuum pact highlights a parallel race: the push to commercialize hybrid quantum-classical computing.[1]

The premise of hybrid compute is that quantum computers are not designed to replace classical supercomputers or AI accelerators.[2]

Instead, they act as highly specialized co-processors. A classical AI model might handle the vast majority of a workload, such as parsing massive datasets, but hand off a specific, intractable optimization bottleneck to the quantum system.[2][4]

"We believe the next phase of enterprise computing will be shaped by bringing quantum, AI, and high-performance computing together," Quantinuum Chief Executive Officer Rajeeb Hazra said in a statement accompanying the announcement.[1]

To understand the actual capability being deployed, it is necessary to look past the marketing language of "the most accurate commercial quantum computer" and examine the hardware itself.

Helios, which Quantinuum launched commercially in November 2025, is a trapped-ion system featuring 98 physical qubits.[4]

In the quantum realm, physical qubits are notoriously fragile and prone to errors caused by environmental noise or temperature fluctuations.

However, Quantinuum has demonstrated the ability to use those 98 physical qubits to create 48 highly stable "logical" qubits, achieving an average two-qubit gate fidelity of 99.921%.[4]

That fidelity threshold is critical because it allows the machine to run deeper, more complex algorithms before errors overwhelm the calculation.

For enterprise cloud architects, the appeal of the Oracle partnership lies in governance and accessibility. Historically, experimenting with quantum hardware required specialized facilities, bespoke contracts, and entirely separate workflows.[3][4]

By placing Helios directly inside an OCI facility, Oracle allows its customers to access the quantum system using the same identity management, billing, and security frameworks they already use for their classical cloud workloads.[1][4]

Mahesh Thiagarajan, Executive Vice President of Oracle Cloud Infrastructure, noted that the arrangement gives developers "a practical and secure way to explore how quantum computing could complement their existing AI and HPC workloads."[5]

The potential applications span several computationally heavy industries. The companies are targeting pharmaceutical firms looking to simulate molecular interactions for drug discovery, financial institutions running complex risk models, and logistics companies solving large-scale routing optimizations.[2]

There is also a significant energy component to the hybrid equation. As classical AI data centers push the limits of regional power grids—often requiring tens of megawatts to operate—the 60-kilowatt power draw of the Helios system offers a glimpse of a more sustainable computational path for specific tasks.[4]

It is important to note, however, that this capability is not shipping to the general public tomorrow. Tuesday's announcement is a strategic pact, with Oracle stating it plans to preview the OCI quantum service to developers in the "coming months."[1][4]

Financial terms of the deal, as well as the exact date for general commercial availability, remain undisclosed. Until the preview launches, the industry will be watching to see if the reality of hybrid quantum-AI compute can match the ambition of its architecture.[2][5]