Nvidia‘s Current Position in AI Processing
Nvidia holds a commanding position in the marketplace for advanced graphics processing units (GPUs) utilized in artificial intelligence (AI) applications. For the computationally demanding process of training AI models, the company is estimated to control over 90% of the market. Even when considering less intensive inference applications—which are used to run existing AI software—estimates place Nvidia’s market share between 60% and 75%. Historically, the company has focused less on central processing units (CPUs) for AI inference, a category currently dominated by Intel and Advanced Micro Devices (AMD), but it is now poised to enter the broader server CPU market.
Expanding Market Scope and Entering the CPU Sector
Although Nvidia has traditionally focused on GPUs due to their higher margins and greater importance in data center and AI processing, this dynamic is rapidly changing. As AI processing increasingly moves to CPUs, and as the total volume of operational and trained applications grows, the company’s incentives to become a major player in the CPU sector are increasing. Consequently, the hardware leader is making strategic moves to become a more diverse provider of compute technologies and services, suggesting rapid expansion into the server CPU space.
New Offerings and Market Projections
The company provided details of this expansion during its fiscal Q2 2027 report, revealing that its Grace family of CPUs had generated $5 billion in sales during the preceding twelve months. Nvidia anticipates that this represents only the beginning of its growth within the category. The company projects that its data center CPU revenue will more than double between this fiscal year and the next, a growth trajectory that will be heavily influenced by the launch of its next-generation processors.
To support this growth, Nvidia is deploying its Vera family of CPUs alongside its top-tier Vera Rubin GPUs. The company has entered full production of its Vera CPUs, creating an integrated hardware solution that pairs these CPUs with the Vera Rubin GPUs and can be sold as a standalone technology. Nvidia claims that its Vera hardware provides efficiency and performance that significantly surpass existing data center CPUs, positioning the company to potentially capture substantial market share from both Intel and AMD.
