HTS feeders swap a crowded trench for a 24/7 cryogenic plant

Writing in Data Center Dynamics, Venkatesh Janakiraman, a professional engineer at design firm LJC, makes the case that high-temperature superconducting feeders have moved past the physics question and into the scheduling question, for reasons that have less to do with zero resistance than with the crowded dirt outside the electrical yard.

An HTS feeder is a medium-voltage cable whose conductor carries very large current only while it remains in a superconducting state, with the conductor sitting inside a cryostat—an insulated, sealed envelope—while a refrigeration plant circulates liquid nitrogen to hold the cable inside its operating window. Janakiraman is careful not to oversell it: HTS is simply a compact conductor and insulation system packaged inside a cryogenic envelope.

The benefit is that HTS can reduce the number of parallel feeders needed to move a given block of power; the cost is adding refrigeration, controls, alarms, and recovery procedures to the electrical chain, and for a campus operator that has never run a cryogenic plant, adoption means becoming a small-scale industrial gas operator overnight inside the most reliability-sensitive facility on earth. The risk profile of the feeder shifts from inert asset to process plant, and owners who size the technology solely on ampacity will miss the instrumentation budget, the training burden, and the recovery drills.

The comparison that matters for an owner is the difference between passive and active infrastructure: a conventional medium-voltage feeder is largely passive once it is in the ground, while an HTS feeder is not, because its electrical behavior depends on a thermal plant that has to run continuously, stay instrumented, and respond correctly when something drifts out of range. Since mission-critical owners already apply FMEA to generators, UPS systems, and cooling plants, HTS simply adds a row to that analysis, and the row is the thermal plant, which is why, as Janakiraman puts it, the technology belongs as much in operations and FMEA discussions as in an ampacity study.

Data centers are revisiting HTS because some projects have run out of easy civil answers. When a campus needs another 50 to 100MW, the classic response is more parallel medium-voltage cable; that works until the corridor is full, the roadway cannot be reopened, the easement is fixed, or the outage window is so constrained that adding another circuit stops being the low-risk option. At that point, the schedule is controlled by underground feeder corridors, electrical yard placement, crossings, utility separation, and the sequence of civil work—long before conductor ampacity gets a vote.

None of this is untested physics: DOE-sponsored cable projects in Albany, Columbus, and on Long Island have demonstrated superconducting power cables operating on live utility systems. The open question for data centers, Janakiraman argues, is no longer whether the technology can move power but whether the operating model fits a 24/7 mission-critical campus, where the shorthand of zero resistance is not very useful in a design review. What matters is the operating envelope—temperature margin, coolant flow, and current all have to stay in their lanes—and when those margins erode, the problem moves quickly from physics to operations.

The framing should determine which owners buy HTS and which keep paying for road closures. HTS does not replace the need for power infrastructure; it substitutes a cryogenic plant for a hundred yards of trench, and the trade pays off only for a campus that has truly run out of civil options, while an owner who treats the cable as a swap-in and leaves the operations manual in the drawer is making a bad trade. The winners will be the operators who run the cryogenic plant the way they already run the electrical room: with alarms, procedures, and a team that has drilled the failure modes.

As this publication has argued, power rights are the binding input in the AI buildout, and scarcity sets the clearing price. HTS sharpens that argument: on the biggest campuses, the binding constraint is the civil path that delivers megawatts, rather than the megawatts alone. But the technology adds a new dependency inside the most reliability-sensitive facilities on earth. The vendors to watch are those publishing operating data along with ampacity curves; the owners to watch are those willing to put “cryogenic plant uptime” into their FMEA. Everything else is marketing.