A quantum computer that only works in a lab is a science project, no matter how good its qubits happen to be.
Diraq’s systems are designed to be deployed wherever they’re needed: on-prem at a facility where new batteries are developed, in situ with the tech controlling a smart energy grid. That means the infrastructure, the route to customers, and understanding of how they will use it have to be developed alongside the technology itself.
For Diraq, the data center is where those considerations start to come together.
That’s why we’re installing a silicon spin quantum computer inside an Equinix data center in Sydney. The system will sit alongside standard servers, with self-contained cryogenic cooling and control electronics, and draw less than 20 kW of power.
But the environment is largely beyond Diraq’s control. And that’s the point. A lab is built to accommodate an experiment. A data center is built to run infrastructure, and infrastructure either fits or it doesn’t.
The Equinix installation allows Diraq to bring the realities of commercial deployment into the development process from the beginning, rather than waiting for those learnings when we hit utility scale.
This work complements our systems engineering approach, and it spans four interconnected areas: physics, infrastructure, distribution, and education. Working with data center partners helps us understand how decisions across those areas come together in practice.
Here’s what that looks like across each of the four areas.
Layer 1: Physics and engineering
A quantum computer isn’t just a chip. It’s a chip, plus the control electronics that drive it, the cryo-CMOS layer that bridges qubit and classical control, the error correction stack, the measurement electronics and software, the networking, and the cooling infrastructure that holds the whole thing at temperature.
The ‘chandelier’ you have likely seen in many photos isn’t the quantum computer either, it is just part of the cooling apparatus. All of this technology is developed through deep physical understanding and requires precision engineering, which starts in the lab, but ultimately has to end with deployment.
In a lab, there is tolerance for chaos and sprawl: cables are everywhere, racks live where they fit, configurations shift between experiments. Commercial deployments do not give you that tolerance. They demand compactness and reliability at the same time, so footprint and power consumption stop being engineering trivia and become procurement constraints that shape the selection of a deployment site. The system must fit somewhere that a customer can access it as part of their compute suite.
Layer 2: Infrastructure integration
For Diraq, that somewhere starts with a data center.
We have written before about why quantum is the next data center transition and about the energy and cooling realities that come with it. The short version is that the architectures most likely to deploy at scale this decade are the ones that can sit alongside GPUs in roughly the same envelope of power, cooling, and space.
That sounds like a technical constraint, but in practice, it’s a commercial one. Data center operators are in the middle of an AI-driven rebuild necessitating denser racks and higher-power deployments, with liquid cooling becoming standard. Quantum must land inside the design assumptions being made now for facilities being built now.
Misalignment isn’t a minor inconvenience that can be overcome. It can mean costly retrofits and delays, or no deployment at all.
This is also why deliberate decisions and the right partnerships are critical. Co-designing classical compute with NVIDIA on NVQLink, aligning manufacturing with GlobalFoundries and imec. None of these choices is incidental. They are what make our systems deployable.
Layer 3: Access and distribution
Even a fully deployable quantum computer is useless to a customer who cannot access it.
This is the layer that often gets the least airtime in technical conversations and the most in commercial ones. How does an end user actually run a workload on a quantum machine? On-prem deployment for a handful of large customers? Cloud access through a hyperscaler, the way AWS Braket has done for the first generation of platforms? A dedicated quantum-as-a-service layer? Some combination, at different points in the maturity curve?
The honest answer is that no one yet knows which channel will dominate. What is clear is that data center operators are the critical enabler in most of these models. They control the physical real estate, the customer relationships, and the integration with the rest of the compute stack. The question of how Diraq captures value in the quantum stack is, in large part, the question of how we partner with the operators who control that distribution.
Layer 4: Customer collaboration
When a technology doesn’t yet exist at the scale needed for end-user value, the most valuable conversations are the ones that help customers and partners build their own mental model of where quantum fits, what it accelerates, and what it doesn’t. The customers who will get the most out of utility-scale quantum in the coming years are the ones who start thinking about quantum-ready workloads now.
That’s why ‘education’ and ‘sales’ aren’t separate functions for us. They are very much embedded in the same technical and commercial discussions at different stages.
Delivering on the potential of quantum computing
Commercializing quantum computing means developing more than the machine itself. The engineering, infrastructure, route to customers, and understanding of how they will use the technology all need to develop alongside it.
That is why working with data center operators early is critical to building utility-scale quantum systems. It brings the realities of commercial deployment directly into the development process, while there is still time for those lessons to shape the systems we are building.
Our work with Equinix is an early step in putting that approach into practice. By operating a quantum system inside a commercial data center, we can learn how quantum integrates with the infrastructure, classical computing, and access models it will ultimately rely on at utility scale.
Commercialization cannot be the step that comes after building the quantum computer. It has to be part of how we build it.
