At the Goldman Sachs Communicopia + Technology conference, Nvidia founder and CEO Jensen Huang outlined why he expects the company to sustain its rapid growth through next year. Huang emphasized Nvidia's central role in the AI industry, highlighting that the company’s products are integral to nearly every AI model and lab, including major players such as Anthropic, OpenAI, and Google.
Huang addressed misconceptions about Nvidia’s hardware, noting that modern GPUs used in AI workloads are far more complex and costly than consumer gaming GPUs. For example, a single AI GPU system can cost around $8.5 million and consume significant power, reflecting the scale of Nvidia's data center offerings.
He pointed to strong sales momentum, mentioning a specific system combining 36 Grace CPUs with 72 Blackwell GPUs that is experiencing 27% month-over-month growth. Nvidia’s revenue guidance projects a 70% year-over-year increase, potentially reaching approximately $680 billion next fiscal year, up from an estimated $400 billion this year.
Huang also described Nvidia’s extensive visibility into the AI infrastructure market, tracking data center developments and power usage globally. This insight comes from close partnerships with cloud providers, OEMs, and AI startups, allowing Nvidia to anticipate demand accurately.
The CEO addressed concerns about Nvidia’s investment strategy in AI companies, clarifying that investments are made only when backed by substantial customer contracts, minimizing risk. He characterized the returns from these investments as highly favorable.
While acknowledging that AI startups currently drive much of the industry's growth and infrastructure spending, Huang remains confident in Nvidia’s foundational position. The company’s broad involvement across the AI supply chain positions it well to benefit from the sector’s continued expansion.
Nvidia’s outlook underscores its role as a critical enabler of AI innovation, although the technology sector’s history suggests that long-term dominance is never guaranteed.