Nvidia’s Vera CPU Signals A Bigger Threat To Intel And AMD
Nvidia’s Vera CPU shifts control of the AI server stack, threatening Intel and AMD’s data center margins as system integration and Arm architecture gain ground.
- Nvidia Vera CPU uses 128 custom Arm cores, built on TSMC’s 3nm process, targeting 2027 shipment.
- Intel’s data-center revenue declined 12% year-over-year as AI spending shifts to GPU-centric systems.
- AMD’s EPYC CPU margins are under pressure from Nvidia’s integrated platform strategy, analysts say.
- Vera unifies memory with Nvidia GPUs, cutting data transfer latency by up to 40% for AI workloads.
- Nvidia’s DGX systems will include Vera as default, creating a proprietary stack that locks out competitors.
Frequently Asked Questions
Nvidia Vera is a custom CPU based on Arm architecture, designed specifically for AI data center servers. It integrates tightly with Nvidia GPUs to optimize performance and reduce latency.
Vera allows Nvidia to offer a complete AI server stack (CPU, GPU, networking), capturing more revenue and reducing the need for Intel or AMD processors. This vertical integration pressures their data center margins.
According to reports, Nvidia plans to ship Vera CPUs as part of DGX systems starting in early 2027.
Vera uses a custom Arm architecture with 128 cores, built on TSMC's 3nm process.
Nvidia wants to control the entire AI server platform, improving performance and power efficiency by tightly coupling CPU and GPU, and capturing more value from the data center market.
Topics
Original source
www.forbes.com
Discussion
Join the discussion
Sign in to post a comment or reply.
No comments yet. Be the first to share your thoughts!