TECH

Oracle bets on a new cloud frontier and prepares integration that could transform AI workloads

Artificial intelligence has turned data centers into massive processing hubs, but some companies are already looking toward what comes next. Oracle and Quantinuum have announced a strategic partnership aiming to bring quantum computing, traditional supercomputing, and AI together within a single infrastructure. This initiative could allow developers to tackle extremely complex problems without needing to install a quantum machine on their own premises.

Oracle and Quantinuum have announced a multi-year strategic partnership to integrate quantum computing capabilities into Oracle Cloud Infrastructure (OCI).

The initiative centers on Helios, Quantinuum’s commercial quantum computer, which is set to become available via a new OCI quantum service. This will enable customers to combine the hardware with traditional high-performance computing (HPC) resources and GPU-based infrastructure.

The goal is not to immediately replace conventional computers with quantum machines; rather, the strategy relies on combining different architectures.

The two companies intend to explore how quantum computing, artificial intelligence, and classical supercomputing can work together to address problems requiring extraordinary processing power.

Potential applications include new material discovery, drug development, logistics, energy, and financial modeling.

Universities and research institutions are also among the groups that could benefit from the platform, primarily because accessing a quantum machine via the cloud eliminates one of the technology's major barriers: the need to own highly specialized equipment.

Helios features 98 physical qubits and prioritizes precision... Commercially launched in November 2025, Helios represents the third generation of quantum computers developed by Quantinuum.

The machine utilizes trapped-ion technology and features 98 physical qubits. The system has already been used in demonstrations involving 48 logical qubits and, according to the company, achieves an average fidelity of 99.921% for two-qubit gates. Precision is especially important because errors represent one of the biggest obstacles in current quantum computing. The more reliable the operations performed by qubits, the greater the potential for executing complex algorithms in a useful way.

Another aspect highlighted by the companies is energy consumption.

According to estimates presented by Quantinuum, a Helios system would consume less than 1% of the energy used by today's leading supercomputers. This does not mean quantum machines will automatically replace supercomputers, but it could make them valuable complementary resources for specific workloads.

The aim is precisely to route each part of a problem to the most suitable hardware.

An application could, for example, use CPUs for certain operations, GPUs for artificial intelligence-related tasks, and a quantum processing unit (QPU) for stages where quantum algorithms offer advantages.

Developers will be able to experiment with quantum hardware without purchasing a machine... Installing a quantum computer is far removed from the conventional experience of adding servers to a data center. These systems require highly specialized infrastructure, technical expertise, and specific operating conditions.

The cloud can mask much of this complexity.

With Helios running on OCI infrastructure, the expectation is that customers will have secure, managed access to the quantum computer without needing to purchase or directly manage the equipment.

The system is set to be integrated with Oracle’s compute, storage, networking, identity, and data services, utilizing access controls and governance mechanisms already familiar to the platform's customers.

Oracle plans to unveil a preview version of the OCI quantum service in the coming months.

One of the goals is to simplify the transition between simulation and execution. Developers will be able to initially test applications in simulated environments and subsequently run them on actual quantum hardware.

The service is also expected to combine Quantinuum’s development tools with support for open frameworks designed for hybrid programming.

This could be particularly important because, at least initially, many commercial quantum computing applications will likely rely on this close collaboration between classical and quantum machines.

The real bet lies in the combination of AI and quantum computing... The partnership also reveals how major tech companies are beginning to envision the next stage of computing infrastructure.

AI has dramatically increased the demand for GPUs, energy, and processing power. Quantum computing offers a radically different architecture and, for certain categories of problems, could complement these systems.

For Mahesh Thiagarajan, Executive Vice President of Oracle Cloud Infrastructure, the integration aims to provide developers with a practical way to explore how quantum resources can complement AI and HPC workloads.

Researchers also see advantages in this approach. Having GPUs and QPUs available within the same environment can reduce operational complexity and allow scientific teams to focus their efforts on experiments rather than managing disparate infrastructures.

However, there is still a considerable gap between making quantum computers available and demonstrating broad commercial advantages in real-world scenarios.

That is precisely why cloud integration can be strategic. Instead of requiring companies to make massive investments in experimental hardware, they can begin testing algorithms, identifying applications, and discovering where the technology truly delivers benefits.

The partnership between Oracle and Quantinuum, therefore, represents more than just the arrival of another cloud service. It points toward a future where CPUs, GPUs, and quantum processors can share the same workload, each executing the part of the problem it was designed to handle.

If this hybrid architecture works as expected, the next transformation in enterprise computing may not come from replacing current machines, but from the arrival of a new type of processor working quietly alongside them.

mundophone