Qualcomm announced on Monday that it will introduce processors for data centers that will connect to Nvidia’s chips and enable artificial intelligence.
Data centers now depend heavily on Nvidia’s graphics processing units (GPUs) to train the massive AI models that drive chatbots and other applications. They are frequently combined with a central processor unit (CPU), which AMD and Intel dominate the market for.
According to Qualcomm, a bespoke CPU that can interface with Nvidia’s GPUs and software will soon be introduced for the data center. With the importance of Nvidia’s semiconductors for AI, access to the company’s infrastructure is essential for any business looking to get a foothold in the data center.
Since Qualcomm’s prior attempts to enter the data center business last decade were unsuccessful, the announcement truly represents a re-entry into the market.
The U.S. tech giant’s work with data center CPUs have benefited greatly by Nuvia’s acquisition by Qualcomm in 2021. Nuvia produces processors based on Arm designs.
In an interview with CNBC on Monday, Cristiano Amon, CEO of Qualcomm, stated, “I think we see a lot of growth happening in this space for decades to come, and we have some technology that can add real value added.” “Therefore, I believe our CPU is highly disruptive.”
Without providing any details, Amon stated that the business will disclose the CPU roadmap and the delivery date “very soon.”
The market for CPUs for data centers is still very competitive. Large cloud computing companies like Microsoft and Amazon already create and implement their own proprietary CPUs. Additionally, AMD and Intel are well-represented.
Amon responded to the rivalry by stating that Qualcomm will have a position in the data center CPU market.
“There will be room for Qualcomm, especially in the data center, as long as… we can build a great product, we can bring innovation, and we can add value with some disruptive technology,” Amon stated. “There will be significant investment in this enormously sizable addressable market for many years to come.”
Along with numerous other U.S. tech companies, Qualcomm last week inked a memorandum of understanding to create data centers with Saudi-based AI firm Humain. The Public Investment Fund of Saudi Arabia will oversee Humain’s operations.
Qualcomm’s entry into the data center is a component of a larger plan to diversify its business, which has historically been dominated by sales of smartphone chips and modems. While other companies are considering in-house CPUs, Qualcomm’s major client, Apple, is beginning to design its own modems.
Mario Morales, group vice president of semiconductors and enabling technologies at the International Data Corporation, said that Qualcomm’s entrance into the data center field is a sensible step to improve its diversification — and that the market is large enough for a new entrant.
“The data center will be the fastest growing segment for the semi market overall over the next five years, and we’ve seen the tremendous growth from Nvidia in the last three years, but companies like Qualcomm haven’t benefited from that growth because they haven’t had a position in the space,” Morales stated.
Since it takes time to develop a product roadmap and start interacting with customers, I believe that this announcement starts to change that. I also anticipate that some of the technology they offer to this particular industry will probably arrive over the next few years.
Under CEO Cristiano Amon, Qualcomm has entered the hitherto Intel-dominated automotive and PC chip markets.
During a presentation at Computex in Taipei, Taiwan, on Monday, Amon stated that more than 85 PC designs using Snapdragon X Series chips—which were initially introduced in 2023—are currently available for purchase or in development. Amon said that at its annual summit in September, the company will reveal a new PC chip.
Qualcomm is marketing its chips as being able to conduct AI functions on a device instead of in the cloud, where they are currently usually done, and as being power efficient. Since data doesn’t need to leave a piece of hardware, on-device AI is said to have several benefits, including speedier AI apps and increased security.
