NVIDIA Expected to Spotlight Vera CPUs at Computex 2026 as Agentic AI Reshapes the Chip Market
As Computex 2026 approaches in Taiwan, attention is turning to how the world’s biggest chipmakers will respond to one of the fastest-growing trends in artificial intelligence: agentic AI computing. Analysts at GF Securities believe NVIDIA will use the event to make a strong case for its Vera CPUs, positioning them as a key part of the next generation of AI infrastructure.
The shift toward agentic AI has sparked fresh debate across the semiconductor industry. After Intel’s first-quarter earnings, discussions intensified around whether CPUs could take on a larger role in AI clusters, potentially changing the balance between CPUs and GPUs in data center systems. While GPUs remain central to modern AI workloads, analysts believe the rise of more autonomous AI systems could increase demand for powerful server CPUs capable of handling complex inference tasks.
According to GF Securities, NVIDIA is expected to highlight the performance benefits of its Arm-based Vera CPUs at Computex 2026. These chips are designed for the agentic AI era, where AI systems do more than generate responses. They can plan, reason, act across tools, and complete tasks with a higher level of independence. That evolution may require a different mix of computing resources, with CPUs playing a larger role alongside accelerators.
Analysts expect NVIDIA to claim that Vera CPUs can deliver up to 1.5 times faster computing performance compared with competing x86 processors from Intel and AMD. They also believe NVIDIA may promote Vera as offering twice the overall performance and up to four times greater rack-level density than rival x86-based solutions. If accurate, those advantages could make Vera a major component in future AI server deployments.
The potential growth opportunity is substantial. GF Securities estimates that NVIDIA could ship around 1.2 million Vera CPU units in fiscal year 2027, rising sharply to 4.2 million units in fiscal year 2028. That forecast reflects growing confidence that AI infrastructure demand will continue expanding beyond GPUs alone, especially as inference workloads become more advanced and widespread.
A key reason for this expected growth is the changing nature of AI inference. Traditional AI inference often focuses on producing outputs from trained models, such as text, images, or recommendations. Agentic AI inference is more demanding because it may involve multi-step reasoning, tool use, memory management, and decision-making across longer workflows. These workloads could increase the need for CPUs that are tightly integrated into AI systems.
At Computex, NVIDIA is expected to focus on how Vera fits into its broader AI computing strategy. The company has already built a dominant position in AI accelerators, but the growing importance of CPUs gives it another opportunity to expand its influence across the data center market. By pairing advanced CPUs with its AI platforms, NVIDIA can offer more complete systems for enterprises, cloud providers, and large-scale AI developers.
Intel is also expected to make important announcements at Computex, though analysts believe its strategy will differ from NVIDIA’s. GF Securities expects Intel to focus on its Wildcat Lake chips, which are reportedly aimed at the entry-level laptop market. These processors could help Intel compete more aggressively against Apple’s lower-cost notebook offerings and defend its position in consumer computing.
While NVIDIA’s Computex focus is likely to center on AI data centers and server CPUs, Intel’s expected Wildcat Lake push would target a different segment of the market. This contrast highlights how major chipmakers are adapting to separate competitive pressures. NVIDIA is trying to expand its role in AI infrastructure, while Intel is working to strengthen its presence across both PCs and servers.
The rise of agentic AI could also significantly expand the total addressable market for server CPUs. GF Securities predicts that agentic AI may eventually account for 30% of overall inference computing. If that happens, the server CPU market could grow dramatically, potentially reaching $211 billion by 2030.
In terms of unit demand, analysts expect the server CPU market tied to these trends to grow from 3.7 million units in 2026 to 16.3 million units in 2028. That rapid expansion would mark a major shift in how AI infrastructure is built, with CPUs becoming increasingly important in supporting next-generation inference workloads.
Computex 2026 may therefore become a pivotal event for the AI hardware industry. NVIDIA is expected to use the show to demonstrate that its Vera CPUs are not simply an add-on to its GPU business, but a strategic product family built for the future of AI computing. If the company can prove strong performance, efficiency, and rack density advantages, Vera could become a major force in AI data centers over the next several years.
The broader message is clear: the AI chip race is no longer just about GPUs. As agentic AI grows, the demand for high-performance CPUs, efficient system design, and tightly integrated hardware platforms is likely to accelerate. NVIDIA’s expected Vera CPU announcements at Computex 2026 could signal the next phase of competition in the global AI infrastructure market.
