
However, such events are becoming increasingly frequent—and are contributing to the formation of a system in which the most critical resource for the global economy is valued, paid for, and ultimately reinvested in dollar-denominated assets.
This concept is not new. Since the 1970s, the pricing of oil in U.S. dollars has increased global demand for that currency and generated export revenues for oil-producing countries, which were often channeled into U.S. markets.
However, the “petrodollar” is merely a model, not a prediction. It demonstrates that when production, payments, and the recycling of assets reinforce one another, an indispensable resource can anchor a currency in global markets.
The “Petrodollar” for AI
AI could shape its own version of this dynamic. Discussions about who will “win” in the AI economy often focus on the race to create the most advanced models.
But the real transformation begins when model performance levels out and companies integrate AI into their day-to-day operations. At that point, just as we talk about dollars per barrel today, we may begin to consider the concept of “dollars per unit of computing power.”
At the heart of this metric lies energy. AI is far from a weightless technology. Data centers convert electrical energy into computing power, for which a fee is charged, and which is currently being reserved years in advance.
When Anthropic signs a 20-year lease agreement with TeraWulf, an infrastructure provider, it’s very similar to an industrial company securing long-term production capacity for itself. This is a bet on the supply side that demand for computing power will persist.
On the demand side, this bet may well pay off. Of course, while many companies are experimenting with AI, only a few have restructured their operations to incorporate its use.
What Will Strengthen the Dollar’s Dominance
However, OpenAI’s new consulting division— Deployment Company (DeployCo)—aims to change this situation by sending engineers to companies to integrate AI into their workflows. Once such systems are implemented, computing power will become a fixed operating expense.
This brings us to the first channel that will strengthen the dollar’s dominance. If the AI supply chain is dominated by U.S.-affiliated companies, and prices are set in dollars, global digital production will be forced to draw on dollar liquidity, and revenue will flow into a financial system that is largely dollar-based.
Unlike oil revenues, these inflows will not accumulate abroad before returning to U.S. markets.
But the invoice is only the tip of the iceberg. Beneath it, an invisible infrastructure is taking shape: a payment system through which agency-based commerce will increasingly flow.
OpenAI’spartnership with Visa, aimed at building this infrastructure, points to a more programmable future. If AI agents begin purchasing services, organizing logistics, and restocking supplies with minimal human intervention, payments must be automated and adapted for machines.
This is a second channel that will reinforce the dollar’s dominance. In particular, dollar-pegged stablecoins can enable the programmable settlements required for agent-driven commerce.
The announcement of Open USD—a dollar-pegged stablecoin backed by more than 140 payment, financial, and crypto companies— — signals that dollar-pegged tokens are already being positioned for this transformation.
These two channels may ultimately converge. The same infrastructure that enables programmatic settlements in agency commerce via dollar-pegged stablecoins could facilitate computational payments just as easily as any others.
Thus, invoicing in dollars will be combined with settlements in stablecoins, transforming a technological dependency into a monetary one.
Are There Alternatives?
In principle, tokenized deposits or central bank digital currencies could perform the same function as dollar-pegged stablecoins.
However, payment systems benefit from early network effects. The question is not whether dollar-pegged stablecoins are the only option, but whether they will be the first to operate effectively on an international scale.
This infrastructure also provides feedback, as demand for programmable dollars translates into demand for reliable assets to back them—specifically, U.S. Treasury bonds. Agent-based commerce has barely begun to take shape, yet stablecoin issuers are already among the largest buyers of Treasury bills.
Taken together, these developments point to the potential emergence of an “energy-computing power-dollar” cycle. Electricity powers data centers; data centers generate computing power; computing power enables the automation of business operations, including payments, which facilitate programmable transactions; and stablecoin reserves flow into U.S. Treasury bonds.
The result will be a self-reinforcing cycle linking AI infrastructure, digital payments, and U.S. financial markets.
It is noteworthy that, apparently, no government is coordinating this process.
While the petrodollar was built on formal agreements, its AI-based successor is taking shape primarily through commercial decisions. Cloud service providers secure land and electricity for themselves. AI companies integrate models into services. Payment consortia are building infrastructure for stablecoins. Stablecoin issuers are purchasing Treasury bonds. Each step makes sense on its own; together, they are quietly reinforcing the dollar’s dominance.
How to Reduce Dependence
This has important implications for policymakers outside the U.S.—especially in countries that have spent years trying to reduce their dependence on the dollar. The ASEAN+3 countries, for example, have sought to settle trade transactions in their national currencies, integrate national payment systems, and pool reserves to protect against dollar shortages.
Although the ASEAN+3 countries cannot prevent the formation of a self-reinforcing dollar cycle, they can limit their dependence on it.
The key to success will be integrating energy, artificial intelligence, and payments into a single strategic agenda. Regional data centers powered by affordable and increasingly green energy could expand local companies’ access to computing resources. And the development of tokenized settlements in local currency for agency trading would reduce dependence on dollar-based systems and ensure transaction transparency for regulators.
The goal for other countries is not complete technological self-sufficiency, which is likely unattainable in the near term. Rather, it is about participating in digital production without accepting a new level of dependence on the dollar as an entry fee.
After all, unlike the petrodollar agreements, the emerging AI system does not offer countries a seat at the negotiating table at any summit.
While politicians and economists debate the future of the dollar’s dominance, AI companies, cloud service providers, and payment networks may already be writing it into the next chapter of the international monetary system’s history.
Those who hope to influence this chapter must act now—otherwise, they risk being left out of the picture.

Chenxu Fu,
economist at the ASEAN+3 Macroeconomic Research Office (AMRO).

Xiangguo Huang,
senior economist at the ASEAN+3 Macroeconomic Research Office (AMRO).
© Project Syndicate, 2026.
www.project-syndicate.org





















