Economy March 21, 2026

Power, Not Just Chips: How China Could Close the AI Compute Gap

A Bernstein analysis finds China’s advantage in electricity and scale could allow it to rival U.S. AI compute by 2035 if massive investment follows

By Avery Klein
Power, Not Just Chips: How China Could Close the AI Compute Gap

A recent Bernstein study frames AI leadership as a function of raw compute capacity, which depends on energy supply, data center buildout and semiconductor efficiency. Today the U.S. holds about 35 zettaFLOPS of AI compute versus China’s roughly 5 zettaFLOPS. China’s strength in electricity generation and rapid capacity additions could enable it to match or even exceed U.S. compute by 2035, but doing so would require nearly $1 trillion in AI data center capital spending and confront persistent semiconductor and export-control constraints.

Key Points

  • AI leadership is framed as a function of compute capacity, which depends on energy supply, data centers and semiconductor efficiency - impacts cloud and data-center operators, utilities and semiconductor firms.
  • Today the U.S. has about 35 zettaFLOPS of AI compute versus China’s roughly 5 zettaFLOPS (about 15% of U.S. level) - relevant for technology and AI infrastructure markets.
  • China’s rapid expansion of electricity generation and planned data-center capex could enable parity or exceedance of U.S. compute by 2035, affecting power, construction and enterprise IT sectors.

A new analysis from Bernstein places compute power - the combination of energy, data-center scale and chip efficiency - at the center of the contest for AI leadership. According to the report, the United States currently maintains a substantial lead in installed AI compute, with about 35 zettaFLOPS compared with China’s roughly 5 zettaFLOPS, or approximately 15% of the U.S. level.

Where China holds a marked structural edge is in electricity. The country already produces more than twice as much electricity as the United States and is expanding generating capacity at a rapid clip - adding in excess of 500 gigawatts a year, a pace the analysis says surpasses the rest of the world combined. That scale of power growth gives China the potential to expand data-center capacity aggressively, even if domestic AI chips are less energy-efficient than their U.S. counterparts.

Bernstein lays out scenarios for how compute parity could evolve. In a base-case path, China could achieve parity with U.S. compute capacity by 2035 if it offsets weaker semiconductor efficiency by deploying substantially more hardware and facilities. Reaching that outcome would, the analysis says, require nearly $1 trillion of capital expenditure directed to AI data centers and a concurrent, rapid expansion in power infrastructure and battery storage.

In a more assertive scenario where power availability is the binding constraint on global AI growth, China could surpass the United States. Bernstein projects that under such conditions China’s aggregate AI compute could exceed U.S. levels by a large margin, potentially reaching more than three times the U.S. compute capacity by 2035.

Despite that potential, the analysis underscores persistent bottlenecks. China currently trails in advanced semiconductor technology: domestic AI chips operate at roughly one quarter the efficiency of U.S. equivalents. Bernstein estimates that efficiency could improve, with the gap narrowing so that Chinese chips approach more than 50% of U.S. efficiency by 2035. However, export controls and restricted access to leading-edge manufacturing equipment remain material risks to that trajectory.

The competition is therefore asymmetric: the United States leads on chips and software, while China’s advantages are concentrated in power generation, manufacturing scale and unit-cost efficiency. If energy proves to be the ultimate limiter for AI scaling, China’s electricity and buildout advantage could become a decisive factor.

Bernstein’s conclusion is straightforward in scope: AI supremacy will be influenced as much by megawatts and data-center investment as by advances in microchips. The analysis highlights a pathway by which China could close the compute gap, but it also identifies significant investment and technology challenges that would need to be resolved to realize that outcome.


Risks

  • China’s domestic AI chips are currently roughly one quarter as efficient as U.S. chips, creating a technology-performance risk that affects semiconductor manufacturers and AI hardware vendors.
  • Export controls and limited access to advanced manufacturing tools pose policy and supply-chain risks that could constrain China’s ability to close the efficiency gap - relevant to foundries, equipment suppliers and national-security sensitive supply chains.
  • Achieving parity or dominance requires nearly $1 trillion in AI data-center capital spending plus rapid expansion of power infrastructure and battery storage, creating execution and financing risks for utilities, data-center developers and investors.

More from Economy

What Would Unlock a Recovery in Home Improvement Demand? Mar 21, 2026 Powell lauds Volcker’s resolve as political pressure intensifies Mar 21, 2026 United Says Higher Oil Linked to Iran Conflict Will Strain Airlines, Plans Capacity Pullbacks Mar 21, 2026 Energy-Driven Cost Pressure Is Making Iron Ore Prices More Responsive to Oil Moves Mar 21, 2026 Wider economic fallout from Strait of Hormuz disruptions threatens supply chains and industrial output Mar 21, 2026