New energy delivery technology could save Google and Microsoft tens of millions of dollars annually; Vertical PD up to 3KA can enable a 10% loss reduction
Traditional power delivery systems are known to struggle to meet the power demands of advanced AI chips. Existing architectures rely on bulky power stages and large capacitor banks that take up significant space on the PCB. This setup results in high power losses and limits efficiency, creating a barrier to supporting the latest AI workloads. This issue is particularly relevant for tech giants like Microsoft and Google, which are investing heavily in AI and data center infrastructure.
Empower semiconductors has introduced a new Crescendo vertical power delivery platform to address these challenges by eliminating the need for capacitor banks and delivering power directly to AI chips. Placing the PD system directly beneath the processor reduces the power losses associated with traditional designs and improves overall efficiency.
The new platform leverages the company’s proprietary FinFast technology to provide scalable, on-demand power for power domains up to 3,000 A, making it well suited to the growing demands of modern data centers.
Just the beginning
“The increasing power demands of AI far exceed the capabilities of today’s lateral flow solutions, both in scale and speed. By developing the Crescendo platform, Empower is enabling generations of new AI processors to achieve their performance goals while running efficiently and coolly,” said Tim Phillips, Founder and CEO of Empower Semiconductor.
“With this introduction we enable an efficient, true vertical power supply for our customers – and this is just the beginning, as the Crescendo platform will enable the integration of the power supply directly into the processor at a total power supply density of more than 5A/mm2which sets Empower apart as a technology leader.”
Empower says its technology provides five times higher power density and more than 10% less power loss, reducing power loss in data centers with 100,000 CPUs by more than 8 MW.