Sungrow prepares solid-state transformers for AI data centers for production in 2027

Chinese Sungrow, one of the world's largest solar inverter manufacturers, announced that the explosive growth of AI computing power makes solid-state transformers (SST) an inevitable standard for data center power supply — the company reported this in investor meeting minutes referenced by 36Kr.

Why AI data centers need SST

According to Sungrow, the explosive growth of AI computing poses fundamentally new challenges to data center electrical architecture, and solid-state transformers (Solid State Transformer, SST) are the optimal infrastructure solution for their further scaling. The key advantage of SST is that it eliminates energy losses at intermediate stages of current conversion: traditional multi-stage transformer schemes lose some power at each transition between voltage levels, and at the energy intensity of modern AI clusters, these losses turn into a notable line item of expenses and extra thermal load.

• Foreign cloud service providers (CSP) already highly appreciate Sungrow's SST products
• The company is negotiating with leading North American cloud providers about joint product specification development
• 2026 — small batches of prototypes and testing with key customers
• 2027 — beginning of serial orders
• 2028 and beyond — transition to large-scale production for the new generation of computing chips

When to expect serial deliveries

The company plans to strengthen its own testing infrastructure to increase customer confidence in the reliability of SST solutions. According to the announced schedule, 2026 will be marked by small-batch samples and field tests with major customers, serial orders will begin in 2027, and from 2028 onwards — transition to large-scale production synchronized with the needs of the new generation of computing chips in increasing capacity.

Who is Sungrow

Sungrow is one of the world's largest manufacturers of solar inverters and energy storage systems, long supplying power electronics for the energy industry. The company's entry into the data center power supply infrastructure market shows how the rapidly growing demand for AI computing power is pushing energy equipment manufacturers to expand their product lines far beyond traditional solar generation.

Energy as the new bottleneck for AI

The capacity of individual AI data centers that hyperscale cloud providers are building now is measured in hundreds of megawatts and approaches gigawatt values — this is an order of magnitude larger than typical data centers of the previous generation. At such scales, even a few percentage points of losses in current conversion turn into tens of megawatts of wasted power and additional heat generation that must be dissipated by the cooling system. This is why power electronics — transformers, uninterruptible power supplies, power distribution systems — are becoming the subject of separate engineering innovations alongside the accelerators themselves.

What this means

Energy infrastructure is becoming as much a bottleneck for AI development as the chips themselves: power electronics manufacturers like Sungrow are reorienting their development toward data center needs, and the horizon for serial SST transformer deliveries — 2027-2028 — essentially coincides with forecasts for further increases in computing capacity by leading cloud providers.