Jensen Huang Says AI Needs "1,000 Times More Power Than We Currently Have." These 3 Industrial Stocks Will Deliver It.
The industry's biggest future bottleneck won't be memory chips but rather a lack of electricity.
Overview
Speaking to a Stanford University computer science class in May, Nvidia (NASDAQ: NVDA) CEO Jensen Huang raised eyebrows by saying, "The amount of energy that we need for [AI] computing is likely probably 1,000 times more than we currently have." He didn't add any additional detail to his mental math and even conceded, "I wouldn't be surprised if we're off by a couple orders of magnitude." If his goal was simply to provide some tangible perspective -- for effect -- on what the energy industry needs to accomplish in the foreseeable future, then mission accomplished.
The projected scope makes sense. Also in May, Goldman Sachs predicted that the power U.S. data centers alone would need would more than double between 2025 and 2027, from 31 gigawatts (GW) to 66 GW. At peak usage, this would push the domestic AI industry's consumption from just over 4% of the nation's total power output to well over 8%. Extending this trend, Bloom Energy predicts artificial intelligence could be using up to 12% of the United States' produced electricity by 2030, and it would still just be getting started.
Details
The same dynamics are taking shape outside of the U.S. as well, of course.
Source
Originally published at www.fool.com.
Related Articles
- I’m managing a former colleague who just can’t keep up with her workload. How do I handle this without being a jerk?
- Americans are living longer than our retirement system can handle — and many of us are unprepared
- Fed keeps rates steady. The big question now: What will Microsoft and Meta say on AI capex?