LONDON – Quantum networking has entered the enterprise era, converging more with AI and high-performance computing (HPC) without the pains of current supply chain troubles.

Insights emerged from the QDA Forum, an annual event organized by the Quantum Datacenter Alliance that counts Cisco, NTT Data, and quantum networking disruptor NU Quantum as founding members.

One panel on quantum processing unit (QPU)-HPC deployments gave an update on the convergence between both compute platforms.

Alexei Bylinskii, head of quantum computing services at QuEra, stressed “hybrid workloads don’t exist” yet. But the interlinks exist, with Oxford Quantum Circuits (OQC) CTO Simon Philips sharing that his firm was working with Slurm and other elements to form a “huge internalizing into what we call the orchestration layer.”

“That then combines to our fabric layer, where I've now got quantum processing units around the world, GPUs around the world, CPUs around the world, and I've built a machine learning pipeline to send maybe just 2% of data to any QPU. ... It requires a whole new level of programming language, and this is where we start leaning heavily into AI workloads,” Phillips revealed.

Both OQC and QuEra, along with players like Finland’s IQM, are pushing commercialization of quantum with their wares as AI amplifies demand for higher compute diversity. While this may raise worries of unconstrained AI agents breaking quantum systems, said systems are not parallel in development with the various agentic tools in classical networks today. QuEra is cautiously playing with the technology in limited fashion, having recently deployed an Anthropic agent to help restore quantum computer laser systems.

Equinix provides the quantum commercial crux

One of the biggest driver in quantum’s commercial sea change is data center giant Equinix, which was represented on the panel by Petrina Steele, global lead for AI, quantum, and edge at the data center giant.

Steele revealed Equinix’s Fabric One networking service touts quantum computing-as-a-service (QCaaS), and later explained customers were looking for “pure access” to vendors on a selective basis.

“So those with deep pockets, you can imagine, will buy their own quantum computer to kick around," Steele said. "I think where we start to get to more useful quantum is where we'll start to see QCaaS really kind of take shape and scale."

Steele added that scaling will likely come unencumbered by the current spate of supply chain issues caused by all things AI. While noting GPU scarcity is an ongoing concern, Steele doesn’t expect it to be a major immediate constraint on quantum build-out

“We’re not talking very large amounts of sort of GPUs that some of the quantum providers are building out. I don't think it's going to be a must because I don't think it's going to be an immediate need," Steele said.

Quantum goes from theoretical to B2B

Equinix sees a confluence taking place between AI, HPC, and quantum, with increased market traction. Steele noted many quantum providers “are working really closely on how they actually converge” and changing up their marketing away from pure-breed quantum.

“Some are looking (at) not calling it quantum, let's call it an AI platform enabled by quantum instead,” Steele said.

This change from the theoretical to the commercial doesn’t come without its challenges. One guest on the QDA panel, Ryousei Takano, principal research manager at Japan’s National Institute of Advanced Industrial Science and Technology, seemed to plea with his tongue-in-cheek for less services in case they overstrain quantum in its current nascent state.

Steele freely admitted to SDxCentral that it took “a lot of effort” to get Equinix's various stakeholders on-board, coming as they do from more engineering and development backgrounds.

“A lot of these quantum providers, they don't do enterprise networking. They are scientists and PhDs and mathematicians, so enterprise networking isn't really their forte. Walking them through what an enterprise consumes and how they consume it, and then how they then engineer for that is quite not a long process, but it's a kind of shared journey,” Steele said.

With academics becoming B2B experts, telecom players and network operators are also changing with the advent of quantum. Steele pointed to Italian telecom provider Sparkle, which holds more than 20 points of presence (PoPs) with Equinix.

“With 20 of those PoPs they can now offer quantum network interconnect as a product,” Steele said. “That's one example of a NSP (network service provider) adding quantum into that sort of offering on top of their existing infrastructure.”