Bernstein names CATL, Sungrow and Chinese chipmakers as picks in U.S.-China AI power race

Bernstein projects China will hit 1,936 zettaFLOPS of AI compute capacity by 2035-more than triple the U.S. forecast-driven by cheap renewable power and a 32% annual rise in AI spending.

Categorized in: AI News General Finance
Published on: Mar 23, 2026
Bernstein names CATL, Sungrow and Chinese chipmakers as picks in U.S.-China AI power race

China's AI Push Hinges on Power Infrastructure, Bernstein Says

Compute power-not just semiconductors-will determine which country wins the AI race between China and the U.S., according to research from Bernstein. The analysis identifies specific stocks positioned to benefit from China's massive infrastructure buildout.

China currently has 5 zettaFLOPS of AI compute capacity. The country added more than 500 gigawatts of power capacity last year. If that pace continues, Bernstein projects China will reach 1,936 zettaFLOPS by 2035-more than three times the U.S. level of 511 zettaFLOPS.

The Energy Equation

Electricity accounts for about 30% of China's total energy mix, above the 21% global average. Renewable power costs one-third of U.S. rates, giving China a structural advantage in scaling data centers.

Bernstein analysts said China will need significant investment in battery and grid infrastructure to support rapid power generation growth. Solar has enabled fast expansion, while nuclear and battery-based storage improve supply stability.

Stock Picks

Bernstein named four Chinese stocks as outperform-rated plays on the trend:

  • CATL (Contemporary Amperex Technology): A major battery supplier for electric vehicle makers. Price target: 530 yuan ($76.96).
  • Sungrow: A solar and energy storage company. Price target: 260 yuan.
  • Cambricon: A domestic semiconductor maker. Price target: 2,000 yuan.
  • Hygon: Another semiconductor company. Price target: 280 yuan.

All four companies are listed on mainland Chinese exchanges.

The Spending Gap

China will increase AI-related spending by 32% annually through 2035, primarily in data centers. U.S. spending in the category will grow only 8% per year, though from a higher base.

Bernstein predicts China's domestic AI chips will improve to 50% of U.S. efficiency by 2035, up from 25% currently. This reflects Beijing's strategy of using more lower-power chips to offset U.S. export restrictions on advanced semiconductors.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)