How ‘Agentic AI’ is reviving classic tech stocks like Dell and Intel
The new phase of “agentic AI” is driving demand for central processing units (CPUs) and computer memory hardware, and that’s resurrecting some older tech stocks that had taken a back seat for several years. Shares of Dell Technologies closed 33% higher on Friday, its best day ever , after it reported surging revenue growth. Intel stock went parabolic in April and early May. Chipmaker AMD soared 74% in April and 46% in May. Computer systems are increasingly being designed for AI robots – or “agents” – not human users. The shift is requiring a greater number of more traditional central processors and memory units relative to newer graphics processing units (GPUs), which drove the first phase of the AI buildout. The new type of design accommodates a processing workflow known as orchestration , which Nvidia CEO Jensen Huang highlighted in a speech in Taiwan over the weekend. “[The] orchestration path is orchestrated by some software. And so this is fundamentally an agent. It deals with short-term memory called working memory [and] long-term memory … and so the memory management system is incredibly important,” Huang said. Analysts said Monday that CPU-driven orchestration could represent a new “computing model.” “Mr. Huang outlined the way agentic AI is driving a ‘new computing model’ – a paradigm shift that will require a significant amount of CPU performance,” Aaron Rakers at Wells Fargo wrote Monday. “The proliferation of agents requires CPU-driven orchestration via software … driven by short-term … and long-term memory.” Now analysts are adjusting their forecasts based on the memory-heavy requirements of the agentic phase of artificial intelligence, and the approaches being taken by hardware makers dealing with component supply bottlenecks. Morgan Stanley analysts on Monday admitted they missed Dell’s growth outlook, for example. “Our memory thesis was wrong as Dell is managing supply and executing better than peers,” analysts led by Erik Woodring wrote Monday, finally upgrading Dell to equal weight from underweight and boosting their price target to $448 — below where the stock was trading Monday. “Dell is getting better access to memory supply (and pricing) than many enterprise peers,” Woodring wrote. Dell’s higher 2027 revenue guidance was the result of greater demand for DRAM and NAND chips, combined with constrained supply, said Katherine Murphy at Goldman Sachs. Supply lags demand “Demand continues to exceed supply (DRAM, NAND, and CPUs are the primary constraints), and the company expects to exit the year with meaningful backlog, setting up for continued growth in F2028,” Murphy wrote Monday. Sustained demand for CPUs is proving one of the hallmarks of the agentic phase of systems design, a frequent point in analyst commentary. “Agent AI demand [is] strong across the CPU ecosystem,” analysts led by Vijay Rakesh at Mizuho wrote Sunday. “[Intel’s] market opportunities remain abundant with leading share in server CPU.” As orchestrated agents have become more important for AI, GPU-focused Nvidia didn’t participate in the recent semiconductor rally led by older hardware makers. Wall Street sees this trend continuing. “NVDA benefits significantly from agentic CPU volume, but due to the much larger size of their GPU business, the CPU impact on NVDA’s overall EPS is modest,” analyst Chris Caso at Wolfe Research wrote late last week. CPU-heavy orchestration at the level of data centers and cloud supercomputing could also have implications for personal electronics. At the Taiwan computing conference over the weekend, Nvidia announced a custom system-on-a-chip design for Windows PCs they call RTX Spark. “RTX Spark powers the world’s first Windows PCs purpose-built for personal agents,” Nvidia said in a release. Analysts at Seaport Research Partners on Monday called the launch of the PC CPU line “the headline event” of the Taiwan conference, but also sounded some skeptical notes. “On paper, this looks promising technically, but we do not expect it to be material to [Nvidia] numbers any time soon,” analyst Jay Goldberg wrote. “We think it will take several generations before these laptops can appeal to a broader audience.”
About the Author
Related
Discover more from InfoVera USA
Subscribe to get the latest posts sent to your email.
