Navitas Semiconductor's 'Claros' deal is aimed at easing power bottlenecks in AI data centers.

Navitas Semiconductor (NASDAQ: NVTS) said it has completed its acquisition of Claros, which expands its power-semiconductor portfolio to cover AI infrastructure from electrical-grid conversion through processor-level power regulation. The deal takes the company beyond gallium nitride (GaN) and silicon carbide (SiC) power devices into integrated voltage regulation for AI processors, as part of its shift toward high-power markets.

During a webinar on AI data-center power requirements, President and CEO Chris Allexandre described the transaction as a defining step in the company's "Navitas 2.0" strategy, which has moved away from consumer markets toward high-power applications. He said Navitas expects mobile revenue to be insignificant by the end of 2026, with nearly all revenue coming from high-power markets.

"Power has become a bottleneck for AI compute," Allexandre said. "Delivering more usable power, more efficiently is an enabler of more compute output."

The company said its focus includes AI data centers, grid and energy infrastructure, high-performance computing, and industrial electrification. Its portfolio now combines GaN devices for high-frequency power conversion, SiC products for high-voltage applications, and Claros' integrated voltage-regulator (IVR) technology, which delivers power close to processors.

Navitas presented its "grid-to-XPU" architecture as a response to what it calls the AI infrastructure power wall: the cumulative power losses, heat, power-density constraints and material requirements that arise along a data center's electrical path. The company is promoting an approach built on 800-volt DC distribution, fewer conversion stages and power delivery closer to the processor. It says this design reduces energy losses, heat, copper usage and power-density constraints in AI data centers.

Navitas estimates its 2030 serviceable addressable market has grown from $3.5 billion to more than $8 billion, with AI infrastructure representing nearly 85% of that figure. These are market estimates, not revenue guidance. The company said its next priority is customer adoption.