China Positioned as AI Race Winner on Power and Cost Advantages
Summary
- • China has ~400 GW spare power capacity by 2030, three times global data center needs
- • Western Chinese provinces offer electricity at $0.05/kWh vs. up to $0.40/kWh in parts of the U.S.
- • MiniMax and Zhipu AI shares jumped 20%+ after Nvidia CEO praised open-source framework OpenClaw
- • Jefferies has reduced U.S. tech exposure and expects U.S. AI capital spending to peak in 2026
Details
China projected to have 400 GW of spare power capacity by 2030
According to Goldman Sachs, this is roughly three times the power capacity the world will need to meet data center demand by 2030. The U.S. faces aging grid infrastructure and rising power prices in data center hubs like Virginia.
Chinese western provinces offer electricity at $0.05/kWh vs. up to $0.40/kWh in the U.S.
Baidu CFO Henry He cited these figures at a Jefferies Asia Forum, noting that power represents approximately 35% of AI inference costs. This creates a structural and durable cost advantage for Chinese AI operators running large-scale inference workloads.
MiniMax and Zhipu AI shares surged 20%+ after Jensen Huang praised OpenClaw open-source framework
Nvidia CEO Jensen Huang called OpenClaw 'definitely the next ChatGPT' in a CNBC interview, triggering rapid share price movements in Hong Kong. The reaction illustrates how sensitive Chinese AI equities are to Western validation and open-source momentum.
Jefferies strategists argue China is the 'big winner' in AI due to valuation, energy, and adoption advantages
Jefferies global macro strategist Mohit Kumar said the bank has reduced U.S. tech exposure in favor of China. Global head of equity research Chris Wood cited China's energy abundance as decisive and described the U.S. private equity market as suffering from 'financial constipation.'
Baidu Apollo robotaxis operate at break-even at $0.15/mile in Wuhan due to low domestic component costs
Chinese robotaxi and AV operators benefit from affordable LiDAR sensors, cameras, and batteries sourced domestically. U.S. competitors have had to choose between sensor types due to cost constraints, limiting their hardware configurations and competitive edge.
EHang cites domestic supply chain as key to competitive pricing in aerial vehicles
EHang's aerial vehicles rely on domestically sourced batteries, motors, and electronic components that are substantially cheaper than Western equivalents. CFO Conor Yang noted the cost advantage enables highly competitive selling prices — extending China's AI deployment edge from software inference into hardware-dependent autonomous systems like drones and aerial vehicles.
U.S. AI capital spending expected to peak in 2026, with private credit markets under stress
Jefferies' Chris Wood expects U.S. AI capex to peak this year. BlackRock and Morgan Stanley have moved to cap investor withdrawals amid retail outflows tied to the 'AI scare trade,' and private credit markets have been stressed since the bankruptcy of auto parts supplier First Brands Group.
Chinese AI monetization remains uncertain despite structural cost advantages
Intense domestic competition and low price expectations among Chinese consumers could limit revenue and profit for Chinese AI companies even as deployment scales. Structural cost and energy advantages do not automatically translate to financial outperformance or global market leadership.
Stat = quantitative data point, Market Impact = effect on equities or capital markets, Insight = attributed analyst argument or interpretation, Industry Update = operational or sector development
What This Means
The argument that China could structurally outcompete the U.S. in AI is gaining institutional credibility at major investment banks, driven by concrete energy cost differentials and a deep manufacturing base — not just software capability. For AI practitioners, this signals that inference cost economics may increasingly favor deployment in or through Chinese infrastructure. For investors, the thesis carries execution and geopolitical risk: China's AI companies must still prove they can monetize at scale in a fiercely competitive domestic market.