- Navitas Semiconductor (NVTS, Financial) is set to unveil its next-generation power solutions at the "AI Tech Night" in Taipei on May 21, 2025.
- New Power Supply Unit (PSU) features up to 8.5kW capacity and achieves 98% efficiency, addressing critical AI infrastructure needs.
- Innovations include GaNSafe™, GeneSiC™, and IntelliWeave™ technologies, optimizing power efficiency and density for hyperscale data centers.
Navitas Semiconductor (NVTS), a leader in gallium nitride (GaN) and silicon carbide (SiC) technology, is hosting an "AI Tech Night" event in Taipei to showcase its latest innovations in AI data center infrastructure. The event will provide insights into cutting-edge developments, including the unveiling of a next-generation Open Compute Project (OCP) data center power supply unit (PSU) that achieves world-class power density and efficiency.
The new PSU can deliver up to 8.5kW with an impressive 98% efficiency, in compliance with OCP and Open Rack v3 specifications. This marks a significant progression from Navitas' previous 2.7kW CRPS offering, which already doubled power density and reduced energy loss by 30%. The company’s IntelliWeave™ technology, combined with GaNSafe™ and Gen 3-Fast SiC MOSFETs, enables Power Factor Correction (PFC) peak efficiencies of 99.3% and further reduces power losses by 30% compared to existing solutions.
Navitas is also highlighting the world's first mass-produced 650V Bi-Directional GaNFast™ power ICs and IsoFast™ high-speed isolated gate drivers. These innovations support a transition from traditional two-stage to single-stage power topologies, optimizing design and increasing rack space utilization, crucial for the ever-expanding AI data centers where space and efficiency are of paramount concern.
The “AI Tech Night” will feature keynote speeches, demonstrations, and discussions centered on the transformative impact of Navitas' high-power GaN and SiC technologies. This event underscores Navitas' commitment to pioneering power solutions that align with the rapidly growing demands of AI and hyperscale data centers.
