Musk explicitly pointed out that the electricity shortfall for AI chips will reach at least 15 gigawatts by 2027, as chip production capacity grows by 40%-50% annually while power supply increases by only 10%-20%. Google and Anthropic have already leased computing power from SpaceX, primarily because SpaceX took the lead in breaking through constraints by building its own power plants. Meanwhile, Zuckerberg stated that the shortage of "hundreds of thousands or even millions" of skilled jobs required for data center construction is difficult to fill. The biggest obstacle to AI infrastructure development is shifting from computing power to electricity.
The primary obstacle to AI infrastructure development is shifting from computing power to electricity supply.
At the G20 Science and Technology Ministers' Meeting, Musk and Zuckerberg made a rare joint appearance, issuing a unified warning to global policymakers: shortages of skilled labor and insufficient power supply are becoming two core bottlenecks constraining AI development.
Among these, the electricity crisis is particularly urgent—Wall Street Insights articlestates that Musk explicitly pointed out that significant shortages could emerge as early as next year, and by 2027, the power shortfall for AI chips will reach at least 15 gigawatts. Meanwhile, Zuckerberg acknowledged that Meta is currently facing a shortage of skilled technical workers, with hundreds of thousands or even millions of skilled positions required for data center construction remaining difficult to fill.
Analysts believe this warning has direct and profound implications for the market: competition in AI infrastructure investment has extended from chips and data centers to power generation and energy support sectors. Musk revealed that companies such as Google and Anthropic have begun leasing computing capacity from SpaceX, primarily because SpaceX resolved its power supply issues by building its own power plants. This indicates that energy self-sufficiency is becoming a new dimension of competition in the AI sector.
The meeting was held on September 1 in Chapel Hill, North Carolina, focusing on the opportunities and disruptions brought by AI. According to Reuters, both Musk and Zuckerberg called for accelerating the construction of AI data centers. Demis Hassabis, CEO of Google-owned DeepMind, also delivered remarks via video, comparing the potential impact of AI to "ten times that of the Industrial Revolution."
Electricity Crisis: AI Chip Growth Far Outpaces Grid Expansion
Musk issued the most specific warning to date regarding AI energy bottlenecks at the meeting.
Wall Street Insights articlestates that Musk, citing consensus estimates from analysts, indicated that AI chips are expected to face a power shortfall of at least 15 gigawatts by 2027. The fundamental contradiction lies in the severe imbalance in growth rates: AI chip production capacity grows by approximately 40% to 50% annually, while available power supply outside China grows by only about 10% to 20% per year.
"Clearly, things that grow faster will eventually overwhelm those that grow more slowly," Musk stated, describing this structural contradiction.
Musk also revealed that several companies, including Google and Anthropic, are already leasing computing power from SpaceX, precisely because SpaceX resolved its power supply bottleneck by building its own power plants, enabling rapid expansion. He stated that his companies are constructing the necessary power generation facilities alongside their data centers.
Human Resource Bottleneck: Data Center Construction Faces a Shortage of Skilled Workers
Beyond electricity, the shortage of skilled labor is another realistic dilemma currently facing tech giants.
In a video address, Zuckerberg stated that the demand for AI data center construction is substantial, potentially requiring "hundreds of thousands, or even millions," of skilled technical positions in the future. He frankly admitted that Meta is currently struggling to find enough skilled workers to build data centers, emphasizing that this challenge is not a distant risk but an immediate constraint.
This statement, together with Musk’s warning about power supply, outlines the dual bottlenecks hindering the expansion of AI infrastructure: power supply at the hardware level, and talent supply at the software and construction levels. Both point to the same conclusion—the pace of AI development is outstripping the carrying capacity of existing infrastructure systems.
Regulatory Divergence: European Model Criticized by Musk
Beyond infrastructure issues, differences in the regulatory environment have also become a focal point of controversy at this conference.
According to reports, Musk criticized Europe’s overly stringent technology regulations at the conference, arguing that innovation should thrive in a relatively lenient regulatory environment. He advocated that new technologies should be "legal by default" rather than "illegal by default," stating that while European regulations may not stop technological development, they will significantly slow down the pace of innovation.
This stance aligns with the prevailing policy tone in the United States. The day before, Trump criticized opponents of AI data center construction as "backward and poor," reflecting the broader inclination within the U.S. political sphere to accelerate the development of AI infrastructure.
However, the vigorous push by tech giants has not been without resistance within the United States.
Polls show that most Americans oppose the construction of more data centers in their local areas, with primary concerns focused on rising electricity bills and noise pollution. This creates a certain policy paradox in relation to Musk’s warnings about an electricity crisis: addressing the power shortage requires large-scale construction, yet the construction itself is facing public backlash.
Economic Outlook: AI Could Add Over $20 Trillion Annually to the Global Economy
Despite numerous bottlenecks, technology leaders remain highly optimistic about the long-term economic value of AI.
Wall Street Insights articlestates that Musk estimates AI could expand the global economy by 20% to 30%, equivalent to an annual increase of approximately $20 trillion to $30 trillion. He characterized this forecast as a "rough estimate," but also stated that it is a prediction he is "willing to bet heavily on."
Demis Hassabis, CEO of DeepMind, described the potential impact in broader terms, stating that once human-level AI is achieved, its influence could be ten times that of the Industrial Revolution, while emphasizing that countries must promote its development in a responsible manner.