China’s electricity advantage could pressure costly US AI companies, but warnings of an 89% stock crash remain speculative rather than a consensus forecast.
China’s vast electricity network and lower-cost artificial-intelligence ecosystem could place growing pressure on highly valued US technology companies. However, claims that American AI stocks are about to crash by 89% are speculative rather than an established market forecast.
The dramatic figure comes from an OilPrice.com analysis that quoted risk consultant Mehrdad Emadi. He predicted that many surviving US AI companies could eventually trade at approximately 11 cents for every dollar of their current value—a decline of 89%.
Emadi also forecast that between 35% and 50% of the US AI sector’s total valuation could disappear rapidly under a severe downturn. These are the opinions of one analyst, not projections supported by a published financial model or a broad Wall Street consensus.
Why Electricity Matters in the AI Race
Training and operating advanced AI models requires enormous computing capacity, cooling infrastructure and a dependable electricity supply.
China produces more than twice as much electricity as the United States and has invested heavily in renewable generation, nuclear power and ultra-high-voltage transmission lines. Its grid can transport large quantities of power from energy-rich western regions to industrial and population centres in the east.
The United States, by comparison, operates through separate regional grids. New power plants and transmission projects can face lengthy approval processes, equipment shortages and local opposition.
The International Energy Agency expects US data-centre electricity consumption to rise by approximately 240 terawatt-hours between 2024 and 2030. China’s consumption is projected to increase by around 175 terawatt-hours during the same period. Together, the two countries are expected to account for nearly 80% of global data-centre electricity-demand growth.
Berkeley Lab estimates that American data centres could consume between 9.5% and 15.3% of total US electricity by 2030, illustrating the scale of the infrastructure challenge.
Chinese AI Models Compete on Cost
Chinese developers including DeepSeek, Alibaba, Z.AI and MiniMax have focused on efficient models, low prices and open-weight software that businesses can download or customise.
Although Chinese companies remain constrained by US restrictions on advanced chips, they have continued narrowing the performance gap with leading American laboratories. Reuters reported that some Chinese developers are roughly seven months behind the most advanced US groups while operating under much tighter cost limits.
Cheaper Chinese models could pressure the revenue assumptions behind the enormous sums being invested by American cloud providers, chipmakers and AI start-ups.
If customers can obtain comparable results at substantially lower prices, US companies may struggle to earn sufficient returns from their spending on processors, data centres and electricity.
Is an 89% Crash Likely?
There is no reliable evidence that an 89% collapse in US AI stocks is imminent.
America continues to lead in advanced semiconductors, frontier models, private investment, cloud infrastructure and many high-value software markets. China also faces challenges, including limited access to the most powerful chips, underused data centres and intense price competition that can reduce profits.
The IEA expects global data-centre electricity consumption to approximately double by 2030, suggesting that demand for AI infrastructure could remain strong even if individual companies are overvalued.
China’s electricity advantage is therefore a genuine competitive issue, but it does not automatically translate into a market crash. US AI stocks would face their greatest risk if power constraints, heavy borrowing, weak revenue growth and cheaper Chinese competition occurred simultaneously.
The 89% figure should be treated as a high-risk scenario advanced by one commentator—not as a confirmed prediction.
About the Author
