Skip to content
voltsdaily

Sungrow Launches EnerNeo Solid-State Transformer for 800V DC AI Data Centres

Share
AI-generated illustration for: Sungrow Launches EnerNeo Solid-State Transformer for 800V DC AI Data Centres
AI-generated image

Sungrow Power unveiled EnerNeo, its first fully in-house solid-state transformer (SST), at an AIDC product launch held at its Hefei headquarters on July 9, with Microsoft, AWS, Alibaba Cloud, Tencent, ByteDance, Baidu, BDC, GDS, Chindata, Infineon and UL in attendance.

The product is built on an 800V direct current power architecture, which the company describes as the industry consensus solution for AI data centres. That framing matters for buyers of AI power equipment: the hyperscalers and colocation operators present at the launch are the same names setting procurement specifications for high-density compute halls.

EnerNeo is modular, with power ratings freely combinable from 1.5 MW to 4.5 MW to suit retrofit projects, new hyperscale clusters and phased capacity expansion. Each unit uses a 6+3 module redundancy architecture, which Sungrow says doubles mean time between failures and allows the system to run under-saturated.

The company claims 99.999% system availability, 98.5% system efficiency and a power density of 312 kW per square metre. Intelligent module rotation, under which modules take turns idling, raises overall equipment service life by 20%, according to Sungrow.

Serviceability is pitched at data centre operations teams. Medium-voltage modules use a quick plug-and-pull design with maintenance of 15 minutes or less, low-voltage modules support online hot-swap for zero-downtime maintenance, and low-voltage modules weigh 11 kg so a single technician can handle them. Sungrow lists medium and low-voltage partitioning and an integrated vacuum insulation design among the product's first-in-industry features.

Commercial commitments landed on the day. Sungrow signed strategic cooperation and framework procurement agreements with Guangdong Dongyangguang Technology Holding and Zhonglian Data Group. It also unveiled a Green Token Joint Innovation Center with Alibaba Cloud's intelligent computing group, aimed at AIDC power supply technology innovation.

The SST is not a new line of work for the company. Sungrow began research on core SST technologies in 2015 and in 2018 completed development and grid validation of what it describes as the world's first 35kV/1500V, 6 MW directly connected photovoltaic SST. It now cites more than ten years of SST technology accumulation behind the integrated distribution and power supply product.

A roadmap follows. Sungrow plans UL-certified variants, natively outdoor models and 35 kV SST products. From 2027 to 2030 it intends to release SST products for ultra-fast charging, cascaded energy storage, photovoltaics, wind, hydrogen and the grid. The company says it will widen SST application from compute scenarios to energy scenarios across solar, wind, storage, hydrogen and grid systems.

For an English-language market audience, the read-across is on the equipment layer rather than the site. A solid-state transformer replaces the conventional low-frequency transformer and downstream conversion stages with a power-electronics block, and pushing that block to 800V DC changes where losses and floor space sit in an AI hall. Vendors quoting efficiency at 98.5% and density at 312 kW per square metre are competing on capital cost per megawatt of usable compute power, not on nameplate rating alone.

The attendance list is the other signal. Microsoft, AWS, Alibaba Cloud, Tencent, ByteDance and Baidu are demand-side buyers; GDS, Chindata and BDC are data centre operators; Infineon sits upstream in power semiconductors and UL in certification. Sungrow's stated intention to bring out UL-certified machines points at markets where that certification is the gate to sale.

Solarbe published its account of the launch on July 10, 2026, sourced to Sungrow Power. Sungrow presents EnerNeo as a next-generation power hub for AIDC and as a component of its full energy chain.

Sources

Related