What Changed and Why It Matters

VEIR, backed by Microsoft, is moving superconducting power cables from the grid to inside data centers. The company’s first system carries 3 megawatts of low‑voltage power through liquid‑nitrogen‑cooled cables that it says use 20× less space than copper and extend low‑voltage reach by roughly 5×. Pilots are slated for next year, with commercial availability targeted for 2027. For operators racing to support 200-600 kW racks (and some eyeing 1 MW per rack), this could unlock denser halls and longer low‑voltage runs without adding more electrical rooms.

Key Takeaways

  • VEIR’s cable delivers 3 MW at low voltage with liquid‑nitrogen cooling at -196°C, claiming 20× space savings and 5× longer runs than copper.
  • Impact: Fewer step‑down points, denser pods, and less floor/overhead space consumed by busways and copper bundles.
  • Timing: Pilots begin next year; general rollout in 2027-plan for early design studies, not near‑term standardization.
  • Risks: Cryogenic safety, reliability (quench events), energy overhead for cooling, codes/approvals, and total cost are not yet public.
  • Fit: Greenfield high‑density (≥300 kW/rack) or constrained campuses benefit most; lower‑density colos can likely wait.

Breaking Down the Announcement

VEIR sells a complete system: superconducting cables jacketed to contain liquid nitrogen, cooling hardware, and termination boxes that hand off to conventional copper. The company says it will integrate and manufacture the cable assemblies while sourcing the superconducting materials from established suppliers. At -196°C (–321°F), superconductors conduct with effectively no DC resistance. In practice, AC operation introduces some losses plus the cooling plant overhead. Still, the zero‑resistance conductor eliminates the heat and voltage‑drop that make big copper runs bulky and inefficient at high currents.

Why 3 MW matters: at 480 V three‑phase, that’s roughly 3.6 kA. Today, supporting those currents over distance forces fat copper busways, multiple transformer rooms, and careful voltage‑drop management. VEIR’s premise is straightforward: move the same megawatts through far slimmer cables, run them farther, and keep heat out of the whitespace. The company reports a 20× space reduction versus copper and 5× reach extension, letting operators place step‑downs and switchgear where it’s architecturally optimal rather than where copper forces them.

What This Changes for Data Center Design

For AI‑heavy builds, distribution constraints are becoming the bottleneck. Racks climbing from 200 kW toward 600 kW–1 MW drive massive I²R losses and thermal load in low‑voltage copper. Longer low‑voltage runs without thermal penalties mean fewer electrical rooms, fewer PDUs per MW, and more whitespace reclaimed for revenue‑generating racks. On large campuses, extending low‑voltage reach reduces the number of MV/LV transitions and supports denser clusters without relocating or adding substations.

that said, cryogenics introduces a parallel mechanical system: nitrogen storage or closed‑loop cooling, leak detection, ventilation for oxygen‑deficiency hazards, and maintenance procedures. The net energy picture will hinge on cooling overhead versus avoided copper losses. PUE may improve or degrade depending on implementation; the gain is likely strongest where copper losses and space penalties are steepest (long runs, high currents, retrofits with tight pathways).

Risks, Unknowns, and Compliance

Reliability and safety are the gating factors. Superconductors can “quench” (drop out of the superconducting state) if locally warmed, requiring millisecond detection, coordinated protection, and safe energy dissipation. Fault current coordination with very low impedance conductors must be validated. Cryogenic systems bring oxygen‑deficiency risk and require compliance with NFPA 55 and local mechanical codes, plus NEC/NRTL listings for the cable assemblies and terminations. Operators will want clear SLAs for mean time between service, hot‑swap strategies for cooling components, and demonstrated safe failure modes.

Commercial unknowns remain: capital cost per MW of distribution, operating cost of cryogenic cooling, installation complexity, and vendor lock‑in across the cable, terminations, and cooling stack. The company’s timeline-pilots next year, general availability in 2027-means most buyers should treat this as a strategic option rather than a near‑term standard.

Competitive Context

Alternatives are maturing in parallel. Many hyperscalers are evaluating higher‑voltage distribution deeper into the hall (or medium‑voltage to the rack with SiC power conversion), reducing current and copper mass. Others push 400 VDC distribution to limit conversion losses. Copper busways continue to scale (4–6 kA+), and liquid‑cooled busbars are emerging. VEIR’s edge is space and thermal footprint for extreme current density and distance; competitors’ edge is simplicity, code familiarity, and known reliability. Expect cross‑over: superconducting approaches where space is the constraint; conventional or higher‑voltage designs where standard gear fits.

Operator’s Playbook: What to Do Now

  • Identify candidate pods: ≥300 kW/rack clusters with long low‑voltage runs or constrained overhead/underfloor space. Model copper losses, thermal load, and aisle space to quantify potential gains from 20× smaller conductors and 5× reach.
  • Run TCO studies: include cryogenic energy, maintenance, nitrogen logistics (or closed‑loop plant CAPEX), and avoided copper/busway, PDUs, and electrical room space. Stress‑test PUE sensitivity.
  • De‑risk compliance early: engage AHJs on NEC/NFPA interpretations, oxygen‑deficiency monitoring, egress, and alarm integration. Require NRTL listing plans and a protection/coordination study that covers quench events and fault currents.
  • Pilot with clear exit ramps: negotiate performance metrics (availability, quench rate, repair times), spares strategy, and fall‑back paths to conventional copper or higher‑voltage designs if targets aren’t met.

Bottom line: VEIR’s superconducting distribution won’t replace copper overnight, but for AI‑dense builds where space and thermal limits block scale, it offers a credible new lever. If 2026 pilots validate reliability and TCO, expect selective adoption in 2027+ greenfields and high‑density retrofits.