Enphase Energy announced on October 27 that its IQ Solid-State Transformer (IQ SST) had converted 13.8 kV AC directly to 800 VDC during laboratory testing in Austin, Texas.
The demonstration used 132 power modules, with 44 connected in series on each of three phases. Each module is rated at 5.5 kW, giving the configuration a combined module rating of approximately 726 kW—not a measured system output. Enphase reported stable 800 VDC output during simulated AI computing load changes.
Each module incorporates gallium nitride (GaN) switching, Enphase's fifth-generation Kestrel control ASIC and a proprietary high-frequency transformer.
Enphase is developing a single-stage conversion architecture intended to reduce the number of power conversion stages between medium-voltage distribution and AI computing racks.
Predictive control, sub-millisecond response and module-level redundancy are among its design objectives. The October announcement did not independently verify all these capabilities.
The company previously described a separate 1.25 MW rack design using 342 modules.
As AI rack power densities increase, operators face growing challenges involving conversion losses, cooling and equipment footprint.
Direct medium-voltage AC-to-800 VDC conversion could change where UPS functions, energy storage and backup protection are integrated. However, the need for ride-through and backup power must still be addressed.
Enphase plans further testing at 34.5 kV and megawatt scale in 2026, followed by customer pilots in 2027. Its April development announcement targeted volume shipments in 2028, subject to further validation and commercialization.
The latest results represent a laboratory demonstration rather than a commercially deployed system.
QuantumScape Introduces Solid-State Battery System for AI Data Centers
Rittal Shows 800 VDC Power Sidecar for High-Density AI Data Centers
Vertiv Q3 2026 Revenue Rises 61% as Data Center Infrastructure Demand Accelerates
Enphase Energy — October 27, 2026 demonstration
Enphase Energy — May 4, 2026 technical white paper announcement
