ARM's latest AGI CPU was unveiled a few days ago, and CEO Rene Haas claimed the platform is ready to bite into x86's market share, but Intel isn't optimistic about the processor's success with agentic AI.
Intel's DC Executive Has Implied that ARM's Claims of Being Superior to x86 Offerings Are a Marketing Stunt
The race for server CPUs has become much more competitive in recent years, as agentic AI workloads have driven a massive rise in the importance of general computing, particularly within the infrastructure industry. This has opened up prospects for manufacturers like AMD/Intel, who were trying to get their share of the AI frenzy; however, with ARM's AGI CPU, the DC segment got a new competitor, one that looks disruptive, at least on paper. However, Intel's EVP and GM of Data Center Group, Kevork Kechichian, believes that the ARM AGI isn't a threat to the company at all.
According to a report by The Register, one of ARM's primary focuses with the AGI CPU is to develop a solution that 'tries' to address the constraints of running agentic workloads on an x86 platform. ARM's official, Mohamed Awad, implied that CPUs from Intel and AMD are not designed to be efficient with agentic AI, and in particular, he took a dig at Intel's SMT (Hyper-Threading) architecture, calling it a legacy implementation. ARM did say that their AGI CPU is specifically designed to cater to modern-day AI workloads through a light SIMD mode and superior compute density.
What happens when you do multithreading? You throw two jobs at the same core, that's how they get to a high thread count. The reality is that your I/O and your bandwidth don't double, so you've just moved the bottleneck elsewhere.
- ARM's Mohamed Awad
However, when Intel's executive was asked about ARM's 'anti-SMT' stance, he said it was a mere marketing move to hide the limitations of ARM's cores, since they cannot support SMT at all. Kechichian also explained that NVIDIA's Vera CPU, with its custom Olympus cores, supports SMT, indicating that it is an important feature. And, to counter ARM's claim of having a CPU with light SIMD extensions, Kechichian pointed towards Intel's Clearwater Forest Xeon CPUs, which are the closest to what the firm has to put up against the AGI CPU. Clearwater Forest boasts a higher compute density with its 288 cores per socket, but it lags behind in memory bandwidth per core.
ARM's AGI CPU is also more dominant due to its broader ecosystem control, given that it has already seen adoption from Meta and that ARM cores are widely known for what they have done for NVIDIA's Grace platform. While Intel faces server CPU demand from hyperscalers, it is mainly confined to networking applications for now; however, as agentic AI and workloads like orchestration become more prominent, there is a greater chance the general CPU TAM will increase, thereby benefiting all players in the race.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
