Jensen Huang, the CEO of Nvidia Corporation, announced new developments at the Nvidia GTC conference in San Jose, California, highlighting the company’s move into central processing units (CPUs). This shift occurs amid soaring demand for graphics processing units (GPUs), essential for developing artificial intelligence (AI) applications.
Nvidia has commenced shipping its proprietary CPU, named Vera, primarily to large cloud service providers, aiming to compete against established manufacturers like Advanced Micro Devices (AMD) and Intel. These rival firms have longstanding relationships with major cloud companies, positioning them strongly in the market.
On Tuesday, Nvidia disclosed specifications and performance benchmarks for Vera, necessary for potential customers to assess the new chip’s capabilities. Notably, clients such as OpenAI, Anthropic, and SpaceX received these chips earlier this year. This expansion into CPUs reflects Nvidia’s strategy to create a vertically integrated system, encapsulating the entire computing experience rather than offering isolated components. This approach intends to enhance GPU performance, ensuring Nvidia’s systems remain preferred by AI research institutions despite increasing competition.
Historically, CPUs were the most crucial components in servers. However, the rise of autonomous AI has reignited focus on the CPU’s role, which manages and supplies data to these AI systems. Financial markets have reacted positively, with AMD and Intel stocks significantly outperforming Nvidia over recent months.
Nvidia anticipates the server CPU market could grow to $200 billion, contradicting estimates suggesting it is currently worth approximately $37 billion. While AMD presently dominates enterprise server CPU market share, Nvidia aims to carve out its unique space. The future will depend on the widespread adoption of Vera, as Nvidia strives to unseat its rivals in this evolving landscape.
Test Your Understanding
How much do you know?
