Intel bets on GPUs to take on Nvidia in AI chip race

At a key industry forum, Intel outlined plans to develop graphics processing units as part of its turnaround strategy, marking a significant expansion beyond its traditional CPU-focused business amid rising demand for AI and high-performance computing chips.

Intel has announced plans to begin producing graphics processing units (GPUs), a move that signals a notable shift in strategy as the chipmaker works to restore competitiveness in a market increasingly driven by artificial intelligence and accelerated computing.

The announcement was made by Intel CEO Lip-Bu Tan during the Cisco AI Summit, where he said the company intends to develop GPUs to support workloads such as gaming, data centre operations, and artificial intelligence model training. GPUs, which differ from Intel’s core central processing units (CPUs), are designed to handle parallel computing tasks more efficiently and have become central to modern AI infrastructure.

Tan indicated that Intel’s GPU push is still in its formative phase, with the company planning to shape its roadmap around customer requirements and evolving market needs. The initiative will be led by Kevork Kechichian, executive vice president and general manager of Intel’s data centre group, who joined the company last year as part of a broader effort to strengthen engineering leadership.

Intel has also brought in senior talent to support the effort, including Eric Demers, who joined earlier this year after more than a decade at Qualcomm, where he held senior engineering roles. These appointments underscore Intel’s intent to build deeper expertise as it moves into more specialised chip segments.

Challenging Nvidia’s dominance in accelerated computing

The GPU market is currently dominated by Nvidia, whose chips have become the backbone of AI data centres worldwide. While Nvidia did not invent the GPU, its early and sustained focus on AI acceleration has helped it establish a commanding lead, particularly in training large language models and other complex AI systems.

Intel’s entry into GPU manufacturing places it in more direct competition with Nvidia at a time when demand for AI-capable hardware continues to surge. However, industry observers note that breaking into the GPU ecosystem will require not only strong silicon design but also robust software, developer tools, and long-term customer partnerships.

The move also marks an evolution in Intel’s broader turnaround narrative. When Tan took over as CEO last year, he emphasised consolidation and renewed focus on core strengths. Expanding into GPUs, while still within semiconductors, represents a calculated broadening of Intel’s portfolio to address structural shifts in computing demand.

As global enterprises accelerate AI adoption, Intel appears to be positioning itself to play a larger role in the next phase of high-performance and AI-driven computing.