Nvidia CEO Jensen Huang emphasizes that energy supply and efficiency are critical constraints for the rapidly growing AI industry, as both manufacturing and operating AI technologies require significant power. He highlights Nvidia’s ongoing improvements in energy efficiency and stresses the need for supportive energy policies and continued investment in energy infrastructure to sustain the industry’s growth.

Nvidia CEO Jensen Huang discusses the critical role of energy in the rapidly expanding AI industry. He emphasizes that energy—whether called electricity or power supply—should always be considered a bottleneck for any industry, especially one growing as quickly as artificial intelligence. The manufacturing and operation of AI technologies require significant energy and infrastructure, making energy efficiency a central concern for companies in this space.

Huang highlights Nvidia’s ongoing efforts to improve energy efficiency with each new generation of its products. He notes that from the Hopper to Blackwell architectures, Nvidia achieved a tenfold increase in energy efficiency, and from Blackwell to Rubin, another tenfold improvement. These advancements directly benefit customers by allowing them to generate more AI output, or “tokens,” per watt of power, which in turn boosts their revenues within the constraints of their available power supply.

Despite these improvements, Huang acknowledges that there is never enough electricity to meet the demands of the industry. He points out that every industrial revolution has faced energy constraints, and the current AI-driven revolution is no different. The anxiety around energy availability is a persistent issue for both companies and their customers, as the growth of AI is fundamentally limited by the amount of power that can be supplied.

Huang also touches on the importance of supportive energy policies, particularly in the United States. He credits former President Trump’s pro-energy growth agenda for enabling the expansion of new industries like AI. Without sufficient energy infrastructure and policy support, he suggests, it would be difficult for the industry to grow at its current pace.

In conclusion, Huang stresses the need for continued investment in all forms of energy to support the AI industry’s growth. At the same time, he reiterates the importance of making the most of whatever energy is available by relentlessly pursuing greater energy efficiency in each new generation of technology. This dual approach—expanding energy supply and improving efficiency—will be essential for sustaining the industry’s rapid development.