The metrics used to assess the artificial intelligence rivalry between the United States and China are changing. The Sept. 24 summit between U.S. President Donald Trump and Chinese President Xi Jinping produced no clear breakthrough on AI. A U.S. proposal for mutual notification of major AI incidents also did not emerge as a formal bilateral mechanism.
At the same time, U.S. long-term interest rates have moved above 5%, adding financing costs to the hundreds of billions of dollars being committed to AI data centers, semiconductors and power infrastructure. China, facing continued U.S. restrictions on advanced chips, is expanding domestic AI processors, cloud infrastructure and data centers while pushing AI deployment across industries.
The competition can no longer be measured by model performance alone. The ability to secure computing infrastructure, finance it at sustainable costs and translate AI into productivity and revenue is becoming equally important.
Summit Leaves AI Competition Intact
Trump and Xi discussed technology issues at their Sept. 24 summit, but announced no concrete agreement to limit AI competition or establish a joint safety framework. Before the meeting, Washington had proposed that the two countries notify each other when a major AI incident with potential national-security implications occurs.
U.S. Treasury Secretary Scott Bessent had raised the issue during talks with Chinese Vice Premier He Lifeng, arguing for greater transparency between the two countries.
“Moving from opaque to more transparency between the number one and the number two AI powers in the world is very important,” Bessent said.
China has not publicly accepted the proposal, while Washington continues to treat controls on advanced AI chips separately from discussions over risk management. The emerging structure allows technology competition to continue while leaving open the possibility of limited communication over risks that could affect both countries.
The next test will come in bilateral working-level discussions. Whether the incident-notification proposal returns, a dedicated AI communication channel is established, and U.S. semiconductor controls remain unchanged will provide the next policy signals.
Capital Costs Enter the AI Competition
Financial conditions are becoming another measure of AI competitiveness.
The yield on the 10-year U.S. Treasury has moved above 5%, increasing pressure on the financing of large data centers and power infrastructure. As AI investment becomes more dependent on corporate debt and project finance, higher long-term rates directly affect required returns and project economics.
Capital expenditure by major technology companies has been one of the main indicators of the AI investment cycle. That measure is becoming less sufficient as spending on data centers, power networks and semiconductors rises and external financing expands.
Corporate borrowing costs, project-finance rates, data-center utilization and cash flow from AI services will increasingly need to be considered alongside headline capital expenditure. Two projects of the same size can have very different economics when their financing costs differ.
The question is therefore no longer limited to how much capital a company or country can deploy. The cost of that capital and the ability to convert investment into sustainable revenue are becoming part of the competition.
China’s AI Position Extends Beyond Model Rankings
China is expanding domestic computing capacity in response to U.S. restrictions on advanced semiconductors. Huawei and Alibaba are building across AI chips, cloud services, data centers and large-scale models.
Alibaba recently unveiled its Zhenwu V900 AI accelerator and set a target of expanding global data-center capacity to more than 20 gigawatts by 2032. CEO Eddie Wu has outlined a broader strategy spanning models, semiconductors and computing infrastructure.
Huawei is also becoming a central supplier of China’s AI infrastructure. At the same time, Eric Xu, Huawei’s rotating chairman, has acknowledged that Chinese AI models have not yet reached the level at which the most advanced frontier risks seen in leading U.S. systems become an immediate concern. Huawei has also said domestic demand for AI computing equipment is exceeding its current production capacity.
That combination illustrates both the constraints and the assets of China’s AI sector. Access to the most advanced semiconductors remains restricted, but China is developing domestic computing infrastructure and has a large industrial base across manufacturing, electric vehicles, e-commerce, finance and logistics where AI can be deployed.
The United States retains major strengths in frontier models, advanced semiconductors, global cloud platforms and deep capital markets. China is building domestic computing capacity while pursuing AI deployment across a broad industrial economy.
The Sept. 24 summit did not change the direction of the U.S.-China AI rivalry. It reinforced the likelihood of prolonged technology competition while drawing greater attention to the economic foundations required to sustain it.
The next stage will not be determined solely by which country produces the strongest model. Computing capacity, financing costs, industrial deployment and measurable productivity will increasingly matter.
The central question is who can secure the computing resources required for AI, finance them at sustainable costs and turn AI deployment into productivity and revenue. Those factors are becoming critical measures of the next phase of U.S.-China AI competition.
