New Chip Cuts AI Data Center Power Use by 30%

Artificial intelligence is hungry for electricity, and researchers in China just found a way to put it on a serious diet.

Scientists at Jiufengshan Lab in Hubei province have created a power module that slashes energy waste in AI data centers by 30%. Led by Li Sichao and released in late 2025, the innovation replaces old silicon chips with gallium nitride, a newer semiconductor material that handles electricity far more efficiently.

The numbers tell an incredible story. In a single 1-gigawatt AI data center, these modules could save nearly 300 million kilowatt-hours annually. That translates to roughly 240 million yuan in cost savings every year for operators.

The breakthrough targets a hidden energy drain that most people never think about. Power conversion for processors eats up about 11% of a data center's total electricity. While AI companies race to build bigger and faster systems, this small component has been quietly wasting massive amounts of power.

Gallium nitride changes that equation completely. The material not only cuts power loss by 30% but also shrinks the physical size of modules by the same amount. Better still, production costs drop by about half compared to traditional silicon chips.

The Ripple Effect

This matters far beyond just tech companies saving money. Large AI facilities can consume as much electricity as an entire nuclear power plant. As artificial intelligence grows more powerful and widespread, that energy demand was threatening to become unsustainable.

The Jiufengshan Lab solution shows how smart material science can make AI's explosive growth cleaner and more affordable. For telecom operators, cloud providers, and major data center owners worldwide, this represents a practical path forward.

Every data center that adopts these modules reduces strain on local power grids. Communities near AI facilities won't face the same pressure on their electricity supply. The environmental benefit multiplies across hundreds of data centers globally.

The timing couldn't be better. AI is transforming everything from healthcare to education, but only if we can power it responsibly. This chip proves we don't have to choose between technological progress and energy sustainability.

Innovation at the smallest scale just made the biggest difference.