Intel Corp. today announced that Raja Koduri, the head of its AXG graphics card division, will move to the role of chief architect, part of a broader update to the chipmaker’s organizational structure.

Koduri (pictured) joined Intel in 2017 as chief architect after leading Advanced Micro Devices Inc.’s graphics card business. Last year, Koduri left the chief architect role to become the head of a newly formed Intel division known as the Accelerated Computing Systems and Graphics Group, or AXG for short. The division has been leading the chipmaker’s effort to develop standalone graphics processing units.

Intel is now reorganizing the teams within AXG to “accelerate and scale their impact and drive go-to-market strategies with a unified voice to customers,” the company stated today. The teams tasked with designing consumer GPUs will move to Intel’s client computing group, which develops chips for use in devices such as laptops. The AXG staffers focused on data center GPUs, in turn, will become part of Intel’s data center and artificial intelligence group.

Koduri is returning to the chief architect role to help lead the company’s chip development efforts. “Raja Koduri will return to the Intel Chief Architect role to focus on our growing efforts across CPU, GPU and AI, and accelerating high priority technical programs,” Intel stated.

Intel includes compact integrated GPUs in the central processing units that it offers for consumer devices. However, the company until recently didn’t make standalone graphics cards. This year, Intel introduced two separate lines of standalone graphics cards: One product series is designed for use in consumer devices such as laptops, while the other targets the data center market. 

Intel’s first three consumer GPUs debuted in March under the Arc brand. The chips are designed for use in laptops and include artificial intelligence features that can automatically enhance video games’ graphics. Intel’s consumer GPUs are based on a chip architecture known as Xe HPG.

Xe HPG is also the basis of Intel’s first data center graphics card lineup, which debuted in August. But Intel’s next data center GPU, Ponte Vecchio, will use a different architecture called Xe HPC. The latter architecture includes additional onboard cache and can more quickly carry out matrix multiplications, a type of mathematical operation that AI models use to process data.

Intel recently rebranded Ponte Vecchio as the Max Series and announced that the chip will start shipping next year. The chip is set to roll out alongside Sapphire Rapids, the next iteration of Intel’s flagship server CPU lineup, which is also slated to launch in 2023.

Photo: Intel