Enphase Builds 800 VDC AI Power Modules in Texas

Enphase IQ solid-state transformer 4 kW modules for 800 VDC AI data centers manufactured in Texas

Enphase Energy is now building 4 kW power modules in Arlington, Texas for its IQ Solid-State Transformer, a modular platform designed to convert medium-voltage AC directly to 800 VDC for high-density AI data centers.

According to the company, hundreds of the compact modules can be combined into IQ SST racks with capacities up to 5 MW. The hardware is moving from design into physical manufacturing as the 800 VDC AI power transition accelerates.

4 kW modules scale into multi-megawatt racks

Enphase’s approach breaks a large solid-state transformer into repeatable power modules rather than relying on a few massive conversion blocks. Each module contains a proprietary high-frequency transformer, gallium nitride power switching and the company’s fifth-generation Kestrel control ASIC.

That combination makes the story as much about GaN power semiconductors as traditional electrical infrastructure. It also parallels onsemi’s work on denser AI-rack power electronics.

Direct conversion from medium voltage to 800 VDC

The IQ SST is being developed to accept medium-voltage AC in the 13.8 kV to 34.5 kV range and produce regulated 800 VDC for AI infrastructure. Enphase is targeting sub-millisecond response to fast-changing compute loads.

The broader industry is moving in the same direction. BitcoinVersus.tech has covered Delta’s 800 VDC power and 3 MW cooling platform, Huawei’s grid-interactive 800 VDC architecture and UL’s new 800 VDC certification work.

Kestrel ASIC coordinates the power system

Kestrel is a custom 22 nm control ASIC designed to coordinate switching, sensing, communications, protection and security across the distributed power modules. Fiber-optic control provides the communication path across the isolation boundary while the high-frequency transformer carries power.

Technical documentation describes a distributed architecture intended to preserve redundancy and serviceability while responding rapidly to AI load changes.

Texas manufacturing is part of the scaling strategy

Enphase says the modules share manufacturing process commonality with its latest IQ Microinverters. The company has shipped roughly 89.4 million microinverters, and it wants to apply that high-volume manufacturing model to AI infrastructure.

In a continuing public update stream, the brand has also been discussing the shift toward 800 VDC racks and its AI Infra Summit work. The U.S. manufacturing angle places the program alongside a wider push to localize critical data-center hardware and semiconductor supply chains.

Complete IQ SST racks are planned for assembly in Texas. Full-system demonstrations are expected in November 2026, with customer pilots planned for 2027. Those are future milestones rather than completed deployments.

Power density is reshaping the AI data center

Moving conversion closer to the final DC distribution voltage can eliminate intermediate conversion stages, but commercial performance still has to be demonstrated at full-rack scale. Electrical efficiency, thermal behavior, partial discharge, redundancy and serviceability will matter once hundreds of modules operate together.

The Enphase program sits beside the rapid growth in rack-level infrastructure documented in BitcoinVersus.tech coverage of Giga Computing’s 700 kW AI factory and Vertiv’s 2.3 MW AI cooling system.

Sources: Enphase Energy manufacturing update, Enphase IQ SST technical documentation and AI Infra Summit materials.


Advertisement:

BitcoinVersus.Tech Editor’s Note & Disclaimer:
We volunteer daily to ensure the credibility of the information on this platform is Verifiably True. BitcoinVersus.tech provides news, technical analysis and educational information and does not provide financial or investment advice. Verify technical specifications and operational requirements with the manufacturer before making purchasing, deployment or investment decisions. If you would like to support to help further secure the integrity of our research initiatives, please donate here: 3C9o19EH5HSiwEPyCTmEKzxhNCbo2X6TTb

9 responses to “Enphase Builds 800 VDC AI Power Modules in Texas”

  1. […] is the same reason BitcoinVersus.tech has been tracking the shift toward 800 VDC power architectures, multi-megawatt cooling systems and high-density liquid cooling. The constraint is no longer merely […]

    Like

  2. […] physical layer is becoming increasingly important as the industry moves toward higher-voltage AI power architectures and denser compute. It also explains why networking manufacturers are adding capacity while […]

    Like

  3. […] BitcoinVersus.tech coverage: Enphase 800 VDC modules · Navitas and Wise AI power · Delta 800 VDC power · ABB 800 VDC […]

    Like

  4. […] that convergence across the stack. Delta is pairing 800 VDC power with 3 MW cooling systems, while Enphase is building 800 VDC AI power modules in Texas. LITEON’s DCX investment pushes in the same direction, but through an equity relationship […]

    Like

  5. […] already covered how 800 VDC is moving from concept toward deployable hardware. It has also tracked Enphase’s 800 VDC power modules and Delta’s 800 VDC power-and-cooling architecture. SolarEdge and Infineon are attacking the […]

    Like

  6. […] grow. BitcoinVersus.tech has also covered Vertiv’s expansion of liquid-cooling services and Enphase’s 800 VDC AI power modules, both reflecting the shift from individual server components toward facility-scale electrical and […]

    Like

  7. […] BitcoinVersus.tech coverage: Electrical voltage/current/power · Line-to-line vs line-to-neutral · Enphase 800 VDC · ABB 800 […]

    Like

  8. […] approaches such as solid-state transformers are being tested to collapse conversion stages, while 800 VDC data-center power systems are pushing conversion deeper into the campus and closer to […]

    Like

  9. […] same trend is visible in 800 VDC-to-low-voltage AI rack conversion and in U.S.-built 800 VDC power modules. The AI infrastructure boom is not only a GPU story; it is creating a parallel race in switches, […]

    Like

Leave a Reply