From Petrodollar to Electric RMB, the Era of AI+Stablecoin Opportunities

Author: Charlie Liu

Foreword

Twenty years ago, when I was in middle school, I became fascinated by the topic of China's energy security, which later opened the door to my career in macro investment, global payments, and cryptocurrency.

Today, during this historical opportunity period of the productivity and production relations revolution brought about by AI and Crypto, I unexpectedly returned to the origin of energy and electricity. A New Anchor for Currency In the AI ​​era, electricity has become the new scarce commodity. Whoever can organize electricity and computing power on a larger scale, at a lower cost, and with greater stability will be better positioned to embed their currency into the next-generation payment network. Stablecoins aren't magic; they simply package a country's industrial, energy, and settlement chains into a programmable "interface." When this interface is connected to power plants and computer rooms, the anchor of currency quietly shifts from gold and oil to the kilowatt-hour. If we look at the timeline of the dollar's story: the first half relied on the gold standard, the second half on the oil standard, and later on the "debt standard"—the unparalleled fiscal and treasury bond markets provided ultimate liquidity. This system can't collapse overnight, and in this era of "AI revolutionizing productivity and Crypto revolutionizing production relations," it is no longer the only natural option. When the anchor of currency shifts from physical assets to balance sheets, politics and maturity will seep into pricing. Even if the market disagrees with the most pessimistic long-term narrative, it can see that another parallel pipeline is being laid. The stablecoin flywheel Stablecoins are this new pipeline, moving clearing from cross-border intermediaries and messaging systems to open networks and peer-to-peer settlement. Geopolitics makes abstract issues very concrete—when certain banks are excluded from SWIFT and when card schemes suspend services in certain markets, businesses and sovereigns will naturally seek out pathways that remain unplugged. I'm not making any value judgments here, just discussing physical realities: if your exports can be settled using a track that no one else can unplug, your bargaining power will compound at the speed of network effects. This is also why the news about "RMB stablecoins" appears to be a token, but the underlying technology is actually energy. Over the past decade, China has exported not only equipment and engineering projects overseas, but has also transformed its full-stack "power-computing-utilization" capabilities—power generation, transmission, storage, and data centers—into replicable products. Tokens are merely the user experience for settlement; the company's moat lies in the tangible supply capacity built on electricity and reinforced concrete. This closed loop is already operating in some places as a "tokenless version." The power plant's capital expenditures come from China, the equipment comes from China, and the operations, maintenance, and spare parts come from China. Electricity bills are settled in RMB, and funds are linked through Hong Kong, onshore, and offshore accounts. Think of your wall socket as a cash flow entry point. Electricity bills flow through the local distribution grid and ultimately end up in your RMB account, eliminating the need for USD transfers to reap the value. Adding programmable settlement to this pipeline, stablecoins simply increase the speed, turning financing and risk management into code. Energy and Infrastructure Why must it be energy? Because AI has propelled electricity to the center stage of currency. With the widespread adoption of AI, training and inference have evolved from mathematical problems to electrical ones.

Today's data centers already consume a significant share of global electricity consumption, and model scale and service density are still rising.

Large American tech companies are diving headfirst into "clean and stable baseload power": signing nuclear power contracts, locking in capacity through long-term contracts, and integrating distributed power with energy storage.

This isn't ESG sentiment; it's the physical constraints of traditional energy.

The ceiling of AI is determined by the generator behind the socket. Next comes the more pressing question: Who can build the power plant fastest, on the largest scale, on time, and to the highest quality? The integration of wind turbines, photovoltaics, inverters, transformers, DC transmission, phase modulation, energy storage, cooling, and industrial parks—and delivering them on schedule in an unfamiliar geographic and regulatory environment—tests the resilience of industrial clusters, supply chains, and engineering muscle memory. Over the past decade, the continuous iteration of highway, railway, hydropower, UHV, and various industrial park projects overseas has made the "power-engineering-finance" flywheel increasingly effective. The most direct experience came in 2014, when I was working in macroeconomics at Franklin Templeton. On a business trip to Africa, I saw a newly built highway extending all the way to Nairobi, and Zambia's convention center became a local landmark overnight. The engineering team, faced with the complex terrain, simply viewed it as a schedule management issue. You can question capital efficiency, but it's hard to deny the ability to deliver on time in complex environments, which is precisely the scarcest component of the "power-to-money" closed loop. The capital efficiency of investment may not be textbook "optimal," but capabilities are accumulated over time, creating a moat invisible on the books. Oil is certainly still on the scene, especially in the Middle East—a region embracing both crypto assets and new settlement experiments. However, the focus of energy over the next decade will shift even more towards renewable and locally sourced clean electricity. Hydropower, wind and solar power, and energy storage lock value into geography. Coupled with national data sovereignty requirements, local data centers and local electricity are a natural pairing: electricity-to-computing on one side, computing-to-services on the other. Settlement should ideally follow a path that doesn't require traversing other countries' systems.

The Implementation of RMB Stablecoins

There are two very realistic paths.

The first is direct electricity settlement.

Compared to commodity trading, electricity purchase and sales contracts are more compatible with the programmable nature of stablecoins—how much is generated, how much is recorded, how much is paid, with a fully digitized chain, and the currency can follow the meter.

Today, electricity bills, operation and maintenance fees, and financial leases are already denominated in RMB. Tokenization will simply make transactions faster, financing more flexible, and collateral more combinable. The second is settlement for computing power and model services. Electricity becomes "AI output" in the data center, and businesses and developers use stablecoins to purchase API calls, model tokens, vector storage, and inference time. Many cross-border digital services in emerging markets have long used US dollar stablecoins as a "dollar substitute." As supply chains and service providers increasingly originate in China, an offshore RMB stablecoin becomes a convenient second option. If this framework still seems abstract, let's return to a case that has been ridiculed. In 2021, when I was in charge of global strategy at Jack Mallers' Strike, I helped El Salvador establish Bitcoin as legal tender. The president at the time proposed using "volcanic geothermal" mining to transform local resources into globally liquid digital assets. The process wasn't perfect, but the direction was right: transforming "geographically specific" natural endowments into tradable units of value through energy and code. AI and stablecoins are industrializing this idea. At the time, "using volcanoes for mining" was considered a joke. Looking back today, using local energy to digitize value seems like a pioneering example of "electricity standard." The Flywheel of the Second Half of Overseas Expansion Back to the main point: from oil to electricity, the key to closing the loop lies in the "recycling" of currency. In the past, when others questioned the use of RMB for energy settlement, the biggest counter-question was, "What can I buy with that RMB?" The traditional answer is offshore RMB pools, dim sum bonds, and Panda bonds. They can be used, but the market is very thin. The new answer is more direct: buy electricity, buy computing, buy equipment, and buy services. Your generator sets, inverters, energy storage systems, complete vehicles, and charging facilities come from China, as do operations, maintenance, and upgrades. The data center's hardware, software, and services come from China. Then, the "recycling" of foreign exchange reserves won't require dollar transfers. What's more, with the wider global expansion of "Made in China," the RMB can buy things that meet almost every need in life and business. We can even abandon white-label production for overseas brands and engage in true high-quality direct supply. Of course, this means that in the second half of our overseas expansion, we need to work harder on branding and storytelling. But at least the purchasing power of currency and the liquidity pool will naturally grow through the accumulation of supply-side assets. A New Game of Competition and Cooperation among Major Powers To put the conclusion even more bluntly: the winner of the stablecoin war won't necessarily be the best-audited or most regulatory-friendly token, but rather the monetary system most tightly coupled with "low prices, stable electricity, and high-density computing power." If China promotes an offshore RMB stablecoin, its true "secret weapon" won't be the token design, but the ability to deliver wind turbines, photovoltaics, transformers, ultra-high voltage (UHV), and data centers globally, all priced in RMB. That would represent a new monetary order, shifting from an "oil anchor" to an "electricity anchor." Of course, this path won't be without its challenges. The expansion of nuclear power and clean "baseload" electricity will be constrained by approvals and supply chains, making it difficult to achieve overnight. And don't underestimate the US's ability to self-repair: If a compliant, dollar-denominated stablecoin framework is successfully implemented, coupled with significant investment in clean, stable electricity, the US dollar's network effects can be further enhanced at the software level. Regarding direction, AI has already established electricity as a new constraint; faced with this constraint, payments will follow the path of lowest cost, and currency will follow. If I could leave one sentence for readers after finishing this article, I would choose this one: AI makes electricity the "primary variable," while crypto and stablecoins simply connect this physical variable directly to the monetary system; whoever can most effectively organize electricity and computing will be better qualified to define the next generation of monetary interfaces.