Nvidia CEO says AI agents are driving massive spikes in GPU demand

The shift to autonomous agents

Nvidia CEO Jensen Huang recently highlighted a major shift in how people interact with artificial intelligence. He noted that the rise of personalized AI agents is causing both token consumption and graphics processing unit (GPU) demand to go "through the roof." Tokens are the basic units of text that AI models process to generate responses.

Unlike simple chatbots that provide one-off answers, agents are designed to think, reason, and iterate. Huang pointed to specific examples like Meta's Muse, the Grok bot, and Gemini as examples of this new generation of tools. These systems are built to solve complex problems by interacting with various tools and environments autonomously.

Because these agents must perform multiple steps to complete a single task, they require a much higher volume of computation. Huang explained that the need for agents to "think" and "iterate" is the primary driver behind the current surge in hardware requirements. This transition marks a move from humans using AI as a tool to AI acting as an independent worker.

How agents consume more power

The technical reason for this surge lies in how agents operate compared to standard human-to-AI interactions. Gene Munster, a managing partner at Deepwater Asset Management, estimates that these agents can consume 10-15x the tokens of a typical chatbot. This disparity exists because agents do not just provide a single output; they perform internal cycles of reasoning.

Data from the OpenRouter platform supports this trend, showing that agents already consume approximately five times more tokens than human users. This happens because the agent must

Market performance and trends

Nvidia's stock has shown significant movement over the past year, advancing by 22.53% in that period. The company has also seen a 19.02% increase year-to-date, though it experienced a slight decline of 2.38% over the last month. These figures reflect the ongoing volatility and growth associated with the AI hardware sector.

Analysis from Benzinga's Edge Stock Rankings suggests a complex outlook for the company's valuation. While the stock maintains a strong trend in the long and medium terms, it currently shows a weak price trend in the short term. Additionally, the stock is currently assigned a poor value score by these rankings.

New tools for AI development

As companies adopt AI coding tools to handle tasks traditionally performed by human developers, new challenges regarding code quality and context have emerged. This shift has led to the rise of specialized software designed to help AI coding agents better understand complex codebases. These tools aim to help organizations maintain control over the rising costs associated with AI-powered development.

One such company, Bito, was founded by former executives from PubMatic to address these specific needs. The company has already raised more than $10 million from both institutional and individual investors. It is currently seeking additional capital through a Wefunder offering to expand its reach in the growing AI software market.

Portfolio diversification strategies

Investors often look for ways to manage risk when dealing with high-growth technology stocks like Nvidia. Because economic cycles shift and no single asset performs well in every environment, many people choose to spread their exposure across different asset classes. This strategy helps capture steady returns and build long-term wealth that is not tied to the fortunes of just one industry.

Diversification can include looking beyond the stock market to include private real estate and private credit. These assets can provide additional income and stability to a portfolio that is heavily weighted toward stocks. By balancing these different types of investments, people can better manage the risks associated with market fluctuations.