NVIDIA Vera CPU Acquired by Perplexity
NVIDIA is competing for market share against Intel and AMD, with AI startup Perplexity confirming the purchase of NVIDIA's latest Vera CPU.
This chip is designed for AI agents and is 1.5 times faster than traditional CPUs in automated coding tasks.
NVIDIA expects the Vera CPU to generate $20 billion in sales this fiscal year, with OpenAI, Anthropic, and Oracle also planning to make purchases.
Source: Public Information
ABAB AI Insight
NVIDIA has previously dominated AI training and inference with GPUs, and the launch of the Vera CPU targets the continuous operation characteristics of AI agents, filling the gap of traditional CPUs in high-intensity tool invocation scenarios; Perplexity, as an AI search company, is the first to adopt it, continuing NVIDIA's expansion from hardware supplier to full-stack AI infrastructure platform.
On the capital path, NVIDIA is entering the CPU market with the Vera CPU, directing funds towards AI-specific computing hardware, motivated by the demand for sustained high-load computing in the agent era, with the expected $20 billion in sales further solidifying its data center revenue; giants like OpenAI plan to purchase to accelerate ecosystem lock-in.
Similar to NVIDIA's heterogeneous computing layout with Grace CPU and GPU, and the impact of AMD's MI series on the GPU market; the current AI hardware competition is transitioning from training GPUs to inference and agent-specific chips, with NVIDIA quickly laying out full-scenario computing leveraging its software ecosystem advantage.
Essentially, this represents industry chain restructuring: traditional CPUs are designed for intermittent human work, while the continuous operation characteristics of AI agents drive the rise of dedicated chips, with the mechanism shifting the computing paradigm from human rhythm to machine rhythm, allowing NVIDIA to achieve full-stack coverage from GPU dominance to CPU + GPU with products like Vera, reinforcing barriers in the data center market.
ABAB News · Cognitive Laws
Humans pause, AI is perpetual.
Dedicated chips consume the general market.
Paradigm shifts, hardware follows.