Intel Targets Nvidia's AI Inference Dominance With New GPU Planned for Year-End Release

The Quick Take

• Intel's data centre unit is targeting a year-end release for a new AI inference GPU designed to compete directly with Nvidia's dominant H100 and B200 accelerator lineup.
• Intel shares have already surged more than 200% year-to-date as investors price in the company's AI infrastructure re-entry strategy.
• The new inference chip targets the fastest-growing segment of AI compute — model serving at scale — where Nvidia currently commands 80%+ market share.

Intel's data centre unit confirmed it is aiming to release a new AI inference GPU by year-end, marking the company's most direct assault yet on Nvidia's dominance in AI accelerator infrastructure. The inference segment — running trained AI models at scale rather than training them from scratch — is becoming the largest and fastest-growing component of AI compute demand as enterprises deploy production AI applications. Intel's shares have rallied more than 200% year-to-date, reflecting investor confidence in the company's strategic pivot toward AI infrastructure after years of competitive pressure from Nvidia and AMD in the data centre GPU market.

For Nvidia, Intel's inference challenge is strategically significant even if it doesn't immediately threaten market share. Nvidia's H100 and B200 accelerators dominate both AI training and inference, and any credible alternative that data centre operators can deploy expands their negotiating leverage against Nvidia's premium pricing. AMD's MI300X has already made inroads in inference workloads, demonstrating that Nvidia's 80%+ market share in AI acceleration is contested. Intel's Gaudi 3 and the incoming inference GPU face the hurdle of software ecosystem compatibility — Nvidia's CUDA dominance creates significant switching costs that pure hardware performance improvements cannot fully offset.

Investors should watch Intel's year-end GPU launch timeline carefully — delays would signal continued execution challenges in its semiconductor roadmap recovery. The key benchmark is whether Intel's inference GPU achieves competitive performance-per-dollar versus Nvidia's current-generation products at the time of release. AMD's MI350 and Nvidia's next-generation Blackwell Ultra architectures will also be in market by year-end, raising the competitive bar Intel must clear. Watch Intel's data centre revenue trajectory in upcoming quarterly results for early evidence that its AI positioning is translating from narrative to customer wins. INTC options positioning around the launch date will reflect institutional confidence in the timeline.

Synthesized from 1 source.