JPMorgan, OQC, AMD Build Colocated Quantum-AI Data Center

Oxford Quantum Computing (OQC), JPMorgan Chase, and AMD have launched a research collaboration centered on a new quantum-AI computing platform in London that physically integrates quantum processors with classical AI infrastructure—a departure from the remote-access model that has defined most quantum computing deployments to date.

The facility, expected to become operational within 12 months, will house OQC's Genesis quantum system alongside AMD-supported AI and high-performance computing resources. JPMorgan Chase will serve as the first dedicated user, with researchers evaluating hybrid quantum-classical applications in portfolio optimization, quantum machine learning, and algorithm development.

Why it matters

The shift from isolated quantum labs to colocated infrastructure addresses a critical bottleneck in hybrid computing: latency. Quantum optimization algorithms often require thousands of exchanges between classical and quantum systems during a single workflow. When those systems are geographically separated, round-trip delays can render algorithms impractical. Colocation enables the low-latency handshake these hybrid workloads demand—and signals that quantum computing is moving from proof-of-concept experiments toward enterprise-grade integration.

The latency problem in hybrid quantum workflows

Hybrid quantum applications such as the Quantum Approximate Optimization Algorithm (QAOA) depend on tight coordination between classical and quantum processors. Classical systems adjust parameters and send them to quantum hardware, which may execute thousands of measurement shots to generate statistical distributions before returning results.

"It's easy for a high volume of round-trips in the optimization loop to create a latency bottleneck if the two machines are geographically separated," Paul Smith-Goodson, vice president and principal analyst at Moor Insights & Strategy, told Data Center Knowledge. Adding AI to the workflow further increases computational demands. "That's one reason the industry is moving away from standalone quantum labs in favor of integrated data centers to ensure the hybrid handshake occurs at a speed needed to make the algorithm viable."

Unlike most quantum deployments today, which provide remote cloud access to standalone systems, the London platform is designed to colocate quantum hardware with AI and HPC resources inside a dedicated environment that meets enterprise security and operational requirements.

From technology push to enterprise pull

Michael Cuthbert, director of the UK's National Quantum Computing Centre, said the project reflects OQC's broader strategy of delivering quantum computing as a service within data center environments. For financial institutions, data sovereignty requirements could make colocated quantum and classical infrastructure particularly attractive.

But Cuthbert noted that the more significant signal may be JPMorgan Chase's willingness to invest in application development rather than infrastructure alone. "What is meaningful is having blue chip end users investing into QC from an applications perspective, providing client pull instead of tech push to the market," Cuthbert told Data Center Knowledge.

Financial services firms have been among the most active enterprise investors in quantum research, drawn by potential applications in risk analysis, simulation, and optimization. "The financial services industry depends on understanding complexity, managing risk, and making decisions with speed, security, and confidence," said Lori Beer, JPMorgan Chase's global chief information officer.

For AMD, the project extends its push to connect emerging quantum systems with the AI and HPC infrastructure already deployed across enterprise environments. "Advancing quantum-AI research will require tightly integrated compute platforms that bring together quantum systems, AI infrastructure, and high-performance classical computing," said AMD CTO Mark Papermaster.

Whether quantum computing delivers near-term business value remains an open question. But the London project highlights how vendors and enterprise users are increasingly thinking about quantum processors as specialized accelerators that will sit alongside GPUs and CPUs in future computing environments—not as standalone systems accessed from afar.

Details of the collaboration were first reported by Data Center Knowledge.