Intel enters the GPU market!
Intel takes on Nvidia! It announced its entry into the GPU market, directly targeting the AI data center business, and has recruited a Qualcomm executive as its chief architect. At the same time, the 14A process foundry has been in deep discussions with clients and is set to ramp up mass production later this year. The market is focused on Intel's subsequent progress in GPU productization and foundry mass production
Intel announced that it will begin manufacturing graphics processing units (GPUs), officially entering the chip market dominated by NVIDIA. This marks a significant strategic expansion for the traditional CPU giant under the leadership of its new CEO, targeting the lucrative data center AI chip business.
According to Reuters, Intel CEO Lip-Bu Tan stated on Tuesday at the Cisco AI Summit that the company plans to manufacture GPUs and has hired Qualcomm senior executive Eric Demmers as the chief architect for GPUs.
Lip-Bu Tan told Reuters that the GPU project will focus on data center scenarios, "tied to the data center business," with the team first collaborating with customers to define product requirements based on that.
Meanwhile, Lip-Bu Tan also mentioned that "several customers" are in deep discussions with Intel's foundry business, Intel Foundry, focusing on the 14A process, with mass production ramp-up potentially starting later this year.
Targeting Data Center GPUs, Directly Facing NVIDIA's Home Turf
According to Reuters and TechCrunch, Lip-Bu Tan clearly pointed the GPU target towards the data center market. Compared to the CPUs long dominated by Intel, GPUs are more oriented towards specific workloads, commonly used for gaming and training artificial intelligence models, which is NVIDIA's core business advantage.
Lip-Bu Tan told Reuters that Intel will first work with customers to define what they need, indicating that its entry strategy will be demand-driven rather than providing a fixed product roadmap upfront.
Key Personnel Move: Eric Demmers Joins, Reporting to Kevork Kechichian
On the personnel front, Intel has filled a key role for the GPU project. Reuters reported that Lip-Bu Tan stated a chief GPU architect has been hired to join the team, and he highly praised their capabilities.
According to TechCrunch and Reuters, Eric Demmers joined Intel in January. TechCrunch noted that Eric Demmers previously worked at Qualcomm for over 13 years, with his most recent position being Senior Vice President of Engineering.
Define Customer Needs Before Product Strategy, Project Still in Early Stages
From Lip-Bu Tan's statements, Intel's GPU plan is still in the strategy and requirement definition phase. TechCrunch reported that Lip-Bu Tan stated the company plans to develop strategies around customer needs, which aligns with his earlier comments to Reuters about "working with customers first, then defining requirements."
This definition approach also extends to capacity and delivery arrangements. Lip-Bu Tan told Reuters that to become a customer, they need to specify the demand and corresponding products so that Intel can plan and allocate time for capacity construction.
Intel Foundry Progressing Simultaneously: 14A Attracting Customers, Mass Production Ramp-Up Later This Year
Outside of GPUs, Lip-Bu Tan also released signals regarding the progress of the foundry business. According to Reuters, he stated that several clients are deeply engaged in cooperation discussions with Intel Foundry, with a focus on the 14A manufacturing technology, and mass production ramp-up may begin later this year.
For the market, this statement connects Intel's two main lines: on one hand, tapping into the incremental demand in data centers through GPUs, and on the other hand, striving for foundry clients with the 14A as a lever. The core concern for the market is whether Intel can simultaneously deliver on GPU productization and foundry mass production progress in response to customer demand
