Elon Musk warned the biggest issue hampering AI advancement in the United States is a problem Chinese competitors don’t have.
In a conversation in Davos, Switzerland, with BlackRock CEO and World Economic Forum interim chair Larry Fink, Musk said AI chip production is increasing exponentially, but electrical power is insufficient, hampering the efficiency of AI data centers in training and deploying AI models.
“I think the limiting factor for AI deployment is fundamentally electrical power,” Musk said. “It’s clear that we’re very soon—maybe even later this year—we’ll be producing more chips than we can turn on.”
The U.S. has been grappling with an outdated grid system, the result of decades of underinvestment and an aging infrastructure. As tech companies increasingly rely on grid operators for electrical power, reliability issues and production limitations have threatened the speed of AI implementation, raising investor concerns of an AI bubble and fueling the belief that the U.S. has already lost the battle with Chinese tech.
Two massive data centers in Nvidia’s Santa Clara, Calif., hometown may sit empty for years waiting for electricity to power them, according to energy experts. Meanwhile, the massive increase in demand, combined with the need for updated infrastructure, have driven up electricity bills for the average American.
Earlier this month, the Trump administration and 13 bipartisan governors mounted pressure on operators of the country’s largest grid, PJM Interconnection, to boost power supply, as well as hold an auction for tech firms to make offers on 15-year contracts to build power plants, which would transfer the cost of electricity away from consumers and to data center operators.
“We know that with the demands of AI and the power and the productivity that comes with that, it’s going to transform every job and every company and every industry,” Interior Secretary Doug Burgum told reporters last week. “But we need to be able to power that in the race that we are in against China.”
During his remarks at the gathering in Davos on Wednesday, President Donald Trump encouraged tech companies to build their own nuclear plants amid the AI push, which he claimed the administration would approve in just three weeks—although these historically take years to approve.
China’s solar edge over the U.S.
Just as many AI investors fear, China is already well ahead of the U.S. when it comes to production capacity, and the country isn’t saddled with the same limitations as the U.S., Musk said at Davos. China is primarily reliant on solar power, seen as a less expensive alternative to nuclear power, with quicker deployment and fewer safety risks.
“China’s growth in electricity is tremendous,” he said.
According to the Global Energy Monitor’s Global Solar Power Tracker, China has nearly four times the amount of operational electricity from solar power than the U.S. Including potential power, China is expected to have 1,118,442 MWac, or electrical power output, from solar energy compared with the U.S.’s 237,947 MWac.
“Solar is by far the biggest source of energy,” Musk said.
Musk claimed powering the U.S. with solar energy would require very little space, only a 100-mile-by-100-mile square of solar fields needed to power the entire country.
But U.S. policies have thwarted efforts to harness and deploy solar power. Despite urging grid operators to take action to increase production capacity, the Trump administration has opposed a pivot to solar energy, stripping subsidies for renewable energy sources it claimed “compromises our electric grid.”
The White House did not immediately respond to Fortune’s request for comment.
Tariffs on solar equipment from Asia took effect in May, with import taxes as lofty as 3,500%, following a U.S. International Trade Commission determination that imports of solar modules and cells from Southeast Asian producers in Malaysia, Thailand, Vietnam, and Cambodia were detrimental to U.S. manufacturers.
“Unfortunately, in the U.S., the tariff barriers for solar are extremely high,” Musk said. “And that makes the economics of deploying solar artificially high.”
