Microinverter Giant Enters AIDC: 4 kW Modules Enter Production, Rack Capacity Reaches 5 MW

<p>In May, we introduced Enphase's IQ solid-state transformer (IQ SST) for AI data centers. At the time, the concept was still at the white-paper stage, with the core idea of using a large number of small power modules to build a megawatt-scale system and convert medium-voltage AC directly to 800 V DC. Four months later, Enphase has begun turning that design into hardware, with plans for a full-system demonstration in November 2026, customer pilots in 2027, and volume shipments targeted for 2028.</p><p>On September 8, Enphase announced that the 4 kW power modules used in the IQ SST have entered production at its Arlington, Texas facility. These modules will be assembled into complete IQ SST racks for megawatt-scale system validation, with the architecture now designed to scale up to 5 MW. The next challenge is to prove that a large number of 4 kW modules can operate reliably as a complete system in a 13.8–34.5 kV medium-voltage environment, while meeting system-level requirements for efficiency, thermal management, insulation, redundancy and maintenance.</p><p><img src="https://energyplanet.oss-cn-shenzhen.aliyuncs.com/uploads/images/20260911174850_47a0cb3f.png" alt="" loading="lazy" decoding="async" /></p><h2>From 4 kW to 5 MW, Enphase is using the same playbook</h2><p>From 4 kW to 5 MW, Enphase is still following the modular approach it knows best. Its earlier white paper showed a 1.25 MW IQ SST rack built from 342 small power modules, while the latest announcement raises the rack capacity ceiling to 5 MW. Although Enphase has not disclosed the exact number of modules required for the 5 MW version or how it relates to the earlier 1.25 MW design, the core logic remains the same: use 4 kW modules as the basic building blocks and scale upward through more modules and system-level coordination.</p><p><img src="https://energyplanet.oss-cn-shenzhen.aliyuncs.com/uploads/images/20260911174850_f21aef63.png" alt="" loading="lazy" decoding="async" /></p><p>This approach closely mirrors the way Enphase has developed microinverters over the past two decades. Traditional high-power conversion equipment typically relies on centralized designs, while Enphase built its microinverter business around breaking a large system into many small, standardized power-electronics units and managing them through a common control layer. It is now applying the same approach to AI data centers: each 4 kW IQ SST module integrates a high-frequency transformer, gallium nitride power switches and Enphase's fifth-generation Kestrel control chip, while software coordinates large numbers of modules to convert medium-voltage AC directly into 800 V DC.</p><p>What makes this approach particularly interesting is that Enphase may be able to reuse not only its product architecture, but also its existing manufacturing system. The 4 kW IQ SST module shares certain similarities in product form and manufacturing processes with the company's latest-generation IQ microinverters. To date, Enphase has shipped around 89.4 million microinverters and has roughly 12 million units of annual microinverter production capacity in the U.S. Large-scale, standardized manufacturing of small power-electronics devices is already one of the company's core capabilities.</p><p>So when Enphase says it ultimately wants to expand IQ SST module production capacity to tens of gigawatts per year and assemble complete racks in Texas, the real question is not the headline number itself, but whether it can carry over the automation equipment, supply chain, standardized manufacturing processes and software-management capabilities already proven in its microinverter business. If those capabilities can be reused at scale, Enphase may face a lower manufacturing barrier than a company developing and mass-producing solid-state transformers from scratch.</p><p><img src="https://energyplanet.oss-cn-shenzhen.aliyuncs.com/uploads/images/20260911174850_20fb4558.png" alt="" loading="lazy" decoding="async" /></p><h2>If the approach works, the role of energy storage could change as well</h2><p>This is where the IQ SST connects most directly with the energy storage industry. Enphase has said the system can achieve sub-millisecond response times to track rapid changes in AI computing loads, particularly sudden rises and falls in GPU power demand. For AI data centers, these fast load swings are a real challenge for front-end power systems, and some architectures use batteries or other short-duration storage near the computing racks to absorb and release instantaneous power.</p><p>If the IQ SST can handle part of those millisecond-scale fluctuations directly at the power-electronics level, some of the work currently performed by rack-side batteries could shift upstream into the conversion equipment itself. But that does not mean storage would be replaced, because the IQ SST addresses power conversion and fast load response, not energy storage. If a data center experiences an actual power outage, or needs backup for several minutes or hours, UPS batteries, site-level BESS, generators and other energy sources would still be required.</p><p>So if the system ultimately reaches commercialization, the more likely change is a redefinition of the boundary between energy storage and power electronics. Millisecond-scale GPU load fluctuations could increasingly be handled by fast power-electronics equipment such as the IQ SST, while UPS backup, peak shaving, grid interaction and longer-duration energy management would remain the responsibility of storage systems.</p><p>The next step for Enphase is still the most basic and difficult one: proving that these 4 kW modules can operate reliably as a 5 MW system while meeting data-center requirements for efficiency, reliability, maintenance and cost. If it succeeds, what Enphase brings into AI data centers will not simply be the design logic of a microinverter, but the broader playbook it has refined over the past two decades: standardized small modules, large-scale manufacturing, and software-based coordination of large numbers of power-electronics units.</p>