Intel Launches GPU Push Into Nvidia-Dominated Sector

Intel has signaled a major shift as it seeks to expand its presence in the graphics processing unit (GPU) market.

Considering this market has long been dominated by Nvidia, Intel’s step is part of a broader industry move to reduce reliance on a single supplier for AI chips.

The announcement highlights growing competitive pressures in the AI semiconductor sector, as rising costs, supply constraints, and geopolitical risks reshape long-standing alliances.

GPU ambitions

Intel CEO Lip-Bu Tan announced a full-scale strengthening of the company’s GPU business during the Cisco AI Summit in San Francisco. The firm has recruited a GPU design chief to assist with that aim.

The company has traditionally dominated the global server central processing unit (CPU) market, but it has struggled to gain traction in high-performance GPUs, where Nvidia controls roughly 80% of the market. Tan said Intel plans to challenge Nvidia’s grip on server GPUs beginning in the second half of this year, using advances in its foundry business, including a planned 1.4-nanometer manufacturing process.

Industry analysts say Intel’s renewed push reflects both technological catch-up and strategic necessity. As AI workloads shift rapidly toward GPU- and accelerator-based computing, reliance on CPUs alone is no longer sufficient for data center operators. Intel’s entry could provide customers with a domestically produced alternative at a time when supply chain resilience has become a key concern for U.S. policymakers.

Cracks in Nvidia’s dominance

Nvidia’s rapid rise has been fueled by the explosive demand for AI training and inference chips. However, as demand surged, prices for high-end GPUs climbed and securing supply became increasingly difficult. These conditions have prompted major technology firms to seek alternatives rather than depend solely on Nvidia’s hardware.

The shift away from Nvidia is most visible among major cloud and platform companies. Google unveiled its seventh-generation AI chip, TPU Ironwood, in November last year and plans to introduce the eighth generation in the second half of this year. Google aims to further enhance the performance of its Gemini AI model using its proprietary Tensor Processing Units.

Microsoft announced its AI inference chip, Maia 200, last month. Manufactured using TSMC’s 3-nanometer process, the chip is designed to optimize high-performance inference workloads. Amazon Web Services launched its energy-efficient custom AI chip, Trainium3, in December, while Meta plans to release its third-generation AI chip, MTIA-v3, in the first half of this year.

Intel’s GPU expansion, aligned with the Trump administration’s push to strengthen domestic semiconductor production, may further accelerate this trend among mid-sized U.S. technology firms seeking alternatives to Nvidia.

Geopolitics and China risk

Geopolitical tensions are also complicating Nvidia’s outlook, particularly in China. Although the U.S. government approved exports of Nvidia’s H200 AI accelerators to China two months ago, shipments remain delayed. The Financial Times reported that Nvidia’s AI chip sales to China are being delayed due to U.S. national security reviews, noting that Chinese customers are postponing orders until conditions become clearer.

Last year, Jensen Huang, Nvidia’s CEO, estimated that annual AI chip sales in China could reach $50 billion in the next couple of years.

Questions over the OpenAI alliance

The market is also watching signs of strain in Nvidia’s relationship with OpenAI, once seen as a core alliance in the AI ecosystem. Media reports, including from The Wall Street Journal, have suggested Nvidia may scale back its investment in OpenAI, while OpenAI has reportedly expressed dissatisfaction with GPU performance. Although both companies’ executives have denied a rift, doubts persist.

Bloomberg reported that Nvidia plans to invest $20 billion in OpenAI’s upcoming funding round, far below the $100 billion figure mentioned last September.

Industry sources say divergence may be inevitable. OpenAI is working with Broadcom to develop its own AI chips and plans to introduce its first custom chip in the second half of this year, alongside a massive 10-gigawatt data center project. “

As Intel enters the GPU arena and major customers chart their own paths, the AI chip market appears set for a more competitive and fragmented future.

Positron has raised $230 million in an oversubscribed Series B funding round, marking one of the largest recent financings in the AI inference hardware space.

The post Intel Launches GPU Push Into Nvidia-Dominated Sector appeared first on eWEEK.