Intel splits AXG graphics unit into two groups under gaming and data center units; AXG's EVP Raja Koduri returns to his previous role as Intel Chief Architect
Intel splits gaming and data center GPUs into two groups. — Intel announced today that it will split its AXG graphics group …
Context & Ripple Effects
Raja Koduri has been the throughline of Intel's graphics push: hired away from AMD in 2017 to build high-end discrete GPUs, then handed the Accelerated Computing Systems and Graphics unit when Intel created AXG in its 2021 reorg. The split announced today unwinds that consolidated structure, folding gaming and data center GPUs into their respective business units.
The move echoes history at both companies: Koduri himself led AMD's 2015 restructuring of its graphics division into a standalone Radeon Technologies Group, so Intel is now reversing toward the org shape AMD abandoned. With Ponte Vecchio and the oneAPI stack anchoring Intel's data center ambitions, the question the split answers is which GPU roadmap gets priority.
First-order effects
- The gaming and data center GPU teams now answer to their respective business-unit leaders rather than a single graphics EVP, meaning separate budgets, roadmaps, and prioritization for Arc versus data center parts.
- Raja Koduri steps out of day-to-day AXG leadership and back into the Intel Chief Architect role he held before taking the graphics unit.
Second-order effects
- Data center GPUs will now compete for capital and engineering attention directly against Intel's core server CPU business instead of enjoying a protected standalone unit — the same tension AMD faced before carving Radeon out in 2015.
- In gaming, Arc now sits inside the client/gaming organization, putting it head-to-head with AMD and Nvidia without a dedicated executive champion at the EVP level.
Third-order effects
- If the pattern holds across the industry, standalone graphics divisions get absorbed into broader accelerated-computing organizations whenever data center demand dominates — a structure that leaves consumer GPUs dependent on spillover from server priorities.
- The coverage shows where this path led at Intel: within months of the split, Koduri left to found a generative-AI startup, underscoring how quickly the dedicated-graphics leadership model gave way once the units were dispersed.
The trend: GPU organizations are being dissolved into wider compute businesses as data center and AI workloads pull graphics strategy out of standalone units and into platform-level priorities.