The tensor contraction bet: Interview with Furiosa.ai CEO June Paik

Furiosa.ai and the architect behind it.

December 10, 2025

David Harold

Furiosa.ai argues that GPUs have the wrong abstraction for AI—and it has spent eight years building a different one. Furiosa.ai, the Seoul-based AI accelerator start-up, has built its data-center inference platform around a tensor contraction abstraction that departs sharply from the matrix-multiplication primitives of GPUs. CEO and co-founder June Paik told us the company’s architecture and compiler were co-designed from Day One to exploit global graph-level optimization rather than thousands of hand-tuned CUDA kernels. With its second-generation RNGD accelerator and a heavily software-weighted engineering team, Furiosa is scaling toward multi-chip parallelism and future training-class systems. In a market where every challenger

...

Enjoy full access with a TechWatch subscription!

TechWatch is the front line of JPR information gathering service, comprising current stories of interest to the graphics industry spanning the core areas of graphics hardware and software, workstations, gaming, and design.

A subscription to TechWatch includes 4 hours of consulting time to be used over the course of the subscription.

Already a subscriber? Login below

The tensor contraction bet: Interview with Furiosa.ai CEO June Paik

Enjoy full access with a TechWatch subscription!

This content is restricted

Login