Navitas Shares Rise 5% on Nvidia AI Power Partnership
Navitas Semiconductor shares rose 5% in after-hours trading following a partnership with Nvidia to build a 24 kW power delivery board for AI data centers.
By Muhamed Porić
October 18, 2026 at 2:00 PM

Navitas Semiconductor (NASDAQ: NVTS) shares climbed 5% in after-hours trading following a Yahoo Finance report on a collaboration with Nvidia to develop a power delivery board for AI data centers. The partnership targets the electrical demands of modern artificial intelligence infrastructure, where high-density computing racks require specialized power conversion architectures to reduce energy loss and thermal output.
Technical Specs of the 24 kW Board
The collaboration centers on a 24-kilowatt power delivery board that directly converts 800 VDC to 12 VDC in a single stage. This design bypasses the traditional 50 V intermediate bus converter stage used in older hardware configurations.
Engineered to fit Nvidia's 400 mm by 60 mm form factor, the new board targets 97% efficiency at full load. Its power density exceeds 2,000 W/in³, reflecting mechanical miniaturization compared to standard industry components.
GaNFast FET Configuration and Topology
The system utilizes 16 of Navitas' NV6034 GaNFast field-effect transistors on the primary side. These components are configured in a stacked half-bridge topology paired with dual-cooled DFN 8 by 8 packaging.
Gallium nitride semiconductors switch faster and handle higher thermal loads than legacy silicon counterparts. This capability allows power supplies to maintain high conversion efficiency even as rack power densities scale upward to support dense clusters of graphics processing units.
Expanding on Prior Data Center Hardware
This 24 kW platform expands Navitas' portfolio of 800 VDC AI data-center solutions, building directly on a previously announced 20 kW board co-developed with Nvidia. The updated hardware is slated to be showcased publicly at the OCP Global Summit 2026.
As major cloud providers and enterprise operators deploy larger GPU clusters, power delivery networks face constraints regarding physical space and cooling capacity. Single-stage conversion topologies help address these constraints by eliminating intermediate voltage steps, reducing both component count and electrical resistance across the power distribution path.
Muhamed Porić
Founder and Editor of Embers.
Newsletter
Get Embers in your inbox
The stories that actually moved something, delivered when there's something worth sending, not daily filler.