Qualcomm It said on Monday it plans to launch processors designed for data centers to power AI, which will link to nvidia'S chip.

NVIDIA's graphics processing unit (GPU) has become a key component in the data center used to train huge AI models to power applications like chatbots. They are usually paired with a central processing unit (CPU), which is a marketplace that Intel and AMD.

Qualcomm said it plans to launch a custom CPU for data centers that can be connected to NVIDIA's GPU and software. For any player who wishes to enter the data center, the link to NVIDIA infrastructure is key given the relevance of the US chip giant to AI.

The announcement actually marks a re-entering Qualcomm's data center market after last year's efforts with a small amount of results.

In 2021, Qualcomm obtained Nuvia, a processor designed based on ARM, and has been the key to the efforts of American technology giants on data center CPUs.

The data center CPU market is still highly competitive. Large cloud computing players like Amazon and Microsoft have designed and deployed their own custom CPUs. AMD and Intel also have strong presences.

Last week, Qualcomm signed a memorandum of understanding with Saudi-based AI Frim Humain to develop data centers and join a series of U.S. technology companies to reach agreements in the region. Howin will operate under the Saudi Arabian public investment fund.

Entering the data center is part of Qualcomm's broader strategy to diversify its business, which has traditionally been dominated by sales of smartphone processors and modems.

Under CEO Cristiano Amon, Qualcomm moved to the bargaining chips for cars and PCs, a market dominated by Intel.

More than 85 PC designs are now selling or developing the Snapdragon X-series chip, first announced in 2023, Amon said in a speech at Computex in Taipei, Taiwan. Amon also said the company will announce new PC chips at its annual summit in September.

Qualcomm pushes its chips toward power efficiency and is able to run AI processes on devices rather than in the cloud, rather than doing it normally in the cloud. Some of the touted advantages of device AI include faster AI applications and more security, as data does not need to leave hardware behind.