Carlton Coffren, a staff scientist at Los Alamos National Laboratory, told the Science, Technology and Telecommunications Committee that data centers for large quantum computers will have different infrastructure needs than the new generation of water‑ and power‑hungry AI centers.

"The data center requirements for large quantum computers are expected to be very different than these large AI data centers," Coffren said, summarizing his briefing that quantum facilities will emphasize cutting‑edge cryogenics, liquid‑helium and liquid‑nitrogen supply chains, vibration isolation and extreme reliability rather than the hundreds of megawatts and million‑gallon‑per‑day water draws some AI centers require.

Coffren showed that currently purchasable quantum systems are room‑scale and generally installable in conventional buildings, though they need a reliable supply of power, liquid nitrogen and helium and careful commissioning: a new machine can take weeks to calibrate and months of continuous operation are preferable to avoid repeating that process. He said today’s devices typically require on the order of kilowatts of continuous power and only small amounts of water as closed‑loop systems, contrasting that with recent examples of AI facilities that have peak demands in the hundreds of megawatts and large daily water needs.

He outlined two broad futures: room‑scale systems that many vendors expect to fit into existing industrial buildings and larger proposals — for example, a multi‑hundred‑million‑dollar Illinois microelectronics park that includes a cryoplant and large cooling infrastructure — for companies that expect greater scale or continuous cryogenic services. Coffren urged state and regional planners to engage with companies now to learn specific requirements and to develop workforce pipelines.

Committee members asked about economic opportunity and workforce. Coffren said New Mexico’s laboratories and universities are strengths for workforce development and that community colleges and regional programs can supply technicians; he recommended talking directly with vendors about 5‑ to 10‑year requirements. He and lawmakers also discussed variability among quantum technologies (superconducting, trapped‑ion/atomic, photonic), and Coffren said it remains too early to predict a single market winner.

Ending: Coffren recommended the committee open early conversations with vendors and labs to understand infrastructure and workforce needs; no legislative action was taken at the hearing.