No Data
Jensen Huang stated that electricity is the biggest bottleneck for AI development, and the solution is to build a large number of small nuclear reactors.
NVIDIA CEO Jensen Huang said on a podcast with Joe Rogan on Wednesday that the future of artificial intelligence will not be constrained by chips, but rather by electricity. He predicted that tech giants will begin to power their data centers with their own nuclear reactors. Huang noted that the rapid expansion of artificial intelligence is running headfirst into a new constraint: electricity. When Rogan asked whether energy had become the biggest obstacle to the development of artificial intelligence, Huang responded without hesitation: 'This is the bottleneck,' emphasizing that the availability of electricity, not GPUs, would determine how far and how fast the industry can scale. Rogan further pointed out that...
Under the Radar: 5 Ultra-Low PE Stocks the Market Is Mispricing
What will be the main AI theme next year? Morgan Stanley: The 'debottlenecking' theme will replace chips, with energy infrastructure seen as a promising sector.
Morgan Stanley stated that with the non-linear growth in AI computing power demand, investors need to shift their focus from 'chips' to 'debottlenecking' by 2026. The firm has once again raised its forecast for data center electricity demand, estimating a cumulative power shortfall of 47 gigawatts for U.S. data centers between 2025 and 2028, based on the latest growth expectations for NVIDIA's sales.
Solar and Nuclear Energy: One Delivers Profits, the Other Gears Up for Growth
Solar energy is a proven profit-generating industry with tangible profits and rapidly growing demand. Nuclear energy, on the other hand, is still in its infancy.
Latest News In Energy Transition - Carbon Credits Drive Growth Amid Climate Goals and Challenges
Macquarie Starts Korea Electric Power at Outperform With KRW65,000 Price Target