Photo: Nattavee/Adobe Stock (Generated with AI)
Quantum computing is moving from laboratory demonstration toward early commercialization, creating a policy challenge that extends beyond funding research or improving quantum bit (qubit) performance. If the United States wants to lead in quantum technologies, it will need to prepare the physical infrastructure that commercialization will require by hardware pathway: Superconducting systems could strain cryogenic and helium-3 supply chains, while neutral atom and trapped-ion systems could strain laser, optics, and vacuum supply chains. And in both cases, physical inputs will determine how quickly quantum computing can scale.
The policy risk is that these infrastructure needs will be addressed only after demand accelerates. AI’s energy bottlenecks show what can happen when infrastructure planning lags deployment, but quantum’s constraints will not be identical. Its most important needs are architecture-specific supply chains, as well as the workforce and regulatory readiness required to operate them. Policymakers still have time to act before these constraints become strategic vulnerabilities.
Because quantum computing remains at an early stage of development, governments, firms, laboratories, and universities have a rare planning window. The difficulty is that quantum does not point to a single infrastructure blueprint: Different architectures impose different demands on power, cooling, materials, and specialized equipment.
These contrasting architectures show why quantum infrastructure cannot be planned around a single technical pathway. Electricity demand will matter, but it is unlikely to be the first or only constraint; the binding inputs will vary by architecture, deployment pathway, and scale.
A recent study by Oak Ridge National Laboratory illustrates why quantum infrastructure planning should extend beyond the electric grid. The study examined operational resource requirements—including electricity, water, nitrogen, helium, and helium-3—across six scenarios for fault-tolerant superconducting quantum computing through 2045. Its central lesson is not that quantum will simply replicate AI’s power demands, but that superconducting quantum systems may be constrained by cryogenic systems, helium-3 availability, cooling infrastructure, and specialized manufacturing capacity. Quantum infrastructure is therefore not only a power problem; it is an architecture-specific industrial supply chain challenge.
Quantum computing also depends on critical raw materials. The materials landscape for qubits spans elemental silicon, III-V semiconductors, diamond, aluminum, and rare earth metals, with different architectures requiring different materials, fabrication processes, and supply chains.
This physical stack makes quantum infrastructure a broader policy issue than the procurement of quantum processors. Enabling commercialization will also require policies that strengthen supply chains and support companies able to manufacture and maintain cryostats, recycle scarce cryogenic gases, build reliable laser and optical systems, secure high-purity materials, train specialized technicians, and connect facilities to sufficient power and cooling capacity.
These requirements suggest that quantum readiness should be treated as an industrial infrastructure challenge, not simply a computing challenge. That distinction matters because quantum’s infrastructure constraints will differ from AI’s, even as AI offers a useful warning about the costs of delayed planning.
The AI comparison is useful not because quantum computing will follow the same demand curve, but because AI shows the consequences of infrastructure planning that trails deployment. Quantum computing’s commercial demand trajectory remains far less certain than AI’s, but even if some demand accelerates, energy systems, permitting processes, supply chains, and capital markets may struggle to adjust quickly enough. They are less mature than those supporting semiconductor fabrication or data center construction today.
The surge in AI data center demand is already forcing the United States to expand energy infrastructure faster than expected. BloombergNEF reported that capital expenditure by the world’s 14 largest data center service companies is expected to reach nearly $750 billion in 2026, up from less than $450 billion in 2025. Of the data centers under construction in the United States as of September 2025, 75 percent are expected to consume more than 23 gigawatts of energy. By 2030, peak energy demand is estimated to add the equivalent of Texas’s entire electricity consumption, about 84 gigawatts, to national demand, which has remained relatively flat since the 1970s. Utilities, grid operators, and federal and state regulators are now responding to a wave of demand that was difficult to anticipate only a few years ago.
AI also illustrates how efficiency gains can increase total resource demand when lower costs enable more use cases, more queries, and broader deployment. This dynamic resembles the Jevons Paradox, originally used to describe coal consumption in nineteenth-century Britain. If quantum systems become more capable and commercially valuable, similar demand dynamics could emerge around specialized facilities, components, and materials even before electricity becomes the dominant constraint.
The lesson for quantum is not to assume an identical demand trajectory, but to plan before constraints harden. Once commercial demand accelerates, scaling energy systems, siting options, grid capacity, specialized facilities, and supply chains becomes far more difficult. The result is not only an economic challenge but also a national security concern.
Lessons from the current AI buildout should shape U.S. quantum policy now. Because quantum computing has not yet reached AI-like scale, the United States still has time to build a more anticipatory infrastructure strategy. Commercialization may accelerate in the early 2030s, but the supply chains, facilities, workforce pipelines, and permitting pathways needed to support deployment will take years to prepare. A preemptive approach should focus on five priorities:
Quantum-ready infrastructure must be understood as more than electricity supply. As quantum computing moves toward commercialization, the United States will need reliable power, cooling-water systems, cryogenic gases, precision components, specialized facilities, resilient supplies of critical materials, and the workforce needed to operate and maintain them. AI’s infrastructure experience reinforces the central lesson of this analysis: When demand accelerates faster than planning, energy systems, permitting processes, capital markets, and supply chains can become strategic constraints.
The United States still has a window to reduce that risk in quantum computing. A quantum-ready infrastructure strategy should map architecture-specific requirements, strengthen vulnerable supply chains, prepare facilities and permitting pathways, build the necessary technical workforce, and use early deployments to collect operational data. These steps will not eliminate uncertainty, but they can reduce the risk that resource bottlenecks might slow commercialization, weaken U.S. leadership, or become strategic vulnerabilities. The question is not whether quantum infrastructure will matter, but whether the United States will prepare before it becomes a constraint.
Hideki Tomoshige is a fellow with Renewing American Innovation at the Center for Strategic and International Studies in Washington, D.C.
The author would like to thank David L McCollum, a distinguished R&D staff member in the Energy Science and Technology Directorate at Oak Ridge National Laboratory, for sharing his insights in preparation for this article.
