Energy diplomacy beneath the competition for compute

The original argument

The CGEP question-and-answer article examines an apparent omission. Despite intense competition over clean technology and AI, the summit’s formal energy discussion centered on diesel and coal rather than solar manufacturing, batteries or electricity for computing. The immediate explanation was the fuel crisis associated with the war in Iran.

The article’s account, in the source account, estimates that Russian and Middle Eastern diesel exports fell from about 1.5 million barrels a day in February to roughly 300,000 in August. US average diesel prices rose from $3.72 a gallon to $5.46 and reached $6.53 in late September. American diesel exports nevertheless reached approximately 1.6 million barrels a day in August. Washington was consequently discussing restrictions on its own exports while asking Beijing to expand refined-product supply.

China did not simply respond to the US request. Its August diesel exports were around 320,000 barrels a day, a relatively high level compared with the period since March 2024, but domestic inventories were near a decade low. Beijing subsequently suspended October refined-product exports. The article treats domestic inventories and supply security as the main determinants of export policy. Since the start of the war, China had been balancing domestic availability, high international refinery margins and relationships with Asian customers.

This creates a distinction between different forms of energy power. The United States can be a leading oil and gas producer while important marginal refining capacity remains in China. Washington sees spare output that could ease global scarcity; Beijing sees an asset to retain against domestic risk.

Coal as an asymmetric bargaining instrument

Coal was included in the proposed “30-for-30” framework, under which each country would purchase about $30 billion of the other’s goods at lower tariff rates. The White House’s account envisaged Chinese purchases of at least 10 million tonnes of US coal in each of 2027 and 2028. The article notes that the Chinese readout did not repeat that quantity, although China’s commerce ministry indicated that inclusion in the framework would help increase imports.

Ten million tonnes would represent only about 2 percent of China’s 2025 coal imports of 492 million tonnes. It could matter much more to American producers. China’s share of US coal exports fell from around 12 percent in 2024 to about 1 percent in 2025, and more than three-quarters of the decline in total US coal exports in 2025 was associated with reduced Chinese purchases.

The same trade therefore has unequal economic and political significance in the two countries. Lower tariffs could improve the competitiveness of American coal and generate political goodwill without materially redirecting China’s energy system. Actual purchases would still depend on relative prices. The unfulfilled energy-purchase targets of the earlier Phase One agreement are a reason to distinguish a diplomatic commitment from assured demand.

The missing electric-vehicle discussion

Expectations of an EV discussion rose around possible attendance by the founders of BYD, Xiaomi and CATL, but they did not attend the state dinner and EVs did not become a formal topic. Chinese vehicles continued to face US tariffs above 100 percent and software restrictions.

Higher gasoline prices nevertheless strengthened the economic appeal of EVs elsewhere. The article records Chinese EV export growth of about 71 percent in January–August 2026 while some US automakers were scaling back domestic EV investment. At the same time, parts of the US industry were considering whether Chinese technology, production efficiency or investment might be necessary to remain globally competitive. Ford’s discussions with the administration and presidential openness to Chinese manufacturing in the United States contrasted with congressional resistance.

The article thus identifies three layers: diesel as immediate crisis supply, coal as a trade-negotiating instrument, and EVs as an unresolved question of industrial competition and cooperation.

The data center connection

The summit did not establish an AI-energy agreement. The connection developed here is an institutional interpretation: competition in chips, models and compute eventually depends on power plants, grids, transformers, fuels and backup systems. Those assets affect the cost and resilience of the physical infrastructure on which AI runs.

China’s computing expansion can draw on a mix of coal, hydro, wind, solar and nuclear generation. Its dependence on imported oil and gas should therefore not be equated directly with the energy exposure of its data centers. Domestic coal and electricity infrastructure can provide a buffer, although transmission constraints, regional conditions and equipment dependencies still matter.

The United States has a different resource position. Abundant natural gas can support an advantage in generation cost and availability, but cheap fuel does not guarantee timely access to a grid connection, transmission capacity or new generating equipment. Resource abundance and deliverable electricity are different things.

From resource diplomacy to institutional capacity

Coal’s role in the negotiations suggests a broader possibility: gas, nuclear fuel, transformers and power projects may also become bargaining instruments within technological competition. This is a prospective research proposition, not a description of an agreement already concluded.

The resulting framework treats national AI capability partly as the ability to convert energy resources into usable computing capacity. The comparison must ask how institutions coordinate investment, secure equipment, authorize connections and distribute risk. A country’s resource endowment is one input; the speed and reliability of that conversion are separate objects of analysis.