Which rung the hardware is on

Start with what exists. North Carolina State University's solid-state transformer is rated at 1 megawatt, it has been sitting in a shipping container at the Electric Power Research Institute's Lenox site in Massachusetts since May 2026, and it returns to the university in September once the summer testing ends. The work behind it goes back to a collaboration with the New York Power Authority that began in 2018.[1]

That is a field demonstration on a live distribution feeder, which is a real rung and a modest one. One unit, one site, one season, and a university research group at the controls rather than a utility carrying load. The 280 million dollars that Amperesand, Heron Power and DG Matrix have raised over the past year belongs on a different rung again: that money is financing, and what financing buys is the attempt to manufacture.[1]

The queue this is meant to escape

The constraint on the other side is well documented and it is old. A conventional power transformer is still built to a nineteenth-century design, wound by hand, and custom-made for each substation, so the largest ones cannot be mass-manufactured and delivery can take several years. This column has watched the same shape of constraint decide a design before: at Abilene, once a turbine backlog became the binding item, waste-heat recovery was the thing given up.[1], [2]

There is a second reason the interest is real, and it is not scarcity. AI halls are moving to direct-current power architectures, and a solid-state unit converts grid alternating current straight into direct current, which removes a separate conversion device from the chain. The plausible alternative explanation is that this architectural fit, more than the delivery queue, is doing most of the pulling, in which case the technology would stay attractive even if conventional transformers arrived tomorrow. The reported evidence supports both readings and separates neither.[1]

The step from demonstration to product

The measurement that would settle it sits in a catalogue entry rather than in another demonstration. If a solid-state transformer supplier publishes a commercial unit with a stated power and voltage rating alongside a quoted delivery lead time by 30 June 2027, the technology has moved from field test to product, and the lead time itself becomes the number to compare against the several years a conventional transformer takes. Until such a figure exists, the honest reading of the Lenox unit is that a substitute for the queue has been demonstrated at 1 megawatt and not yet supplied at any scale.[1]