Will the AI revolution and digital currencies undermine the US dollar hegemony?
In 2025, the financial market staged a drama that left countless seasoned traders stunned.
In this year, against the backdrop of the Federal Reserve maintaining high interest rates and even still exuding the after - effects of monetary tightening, the US Dollar Index had its worst performance during the same period in nearly fifty years.
What makes this event so disturbing is not the decline itself, but the timing. It occurred precisely when all traditional analytical tools pointed to a "should - strengthen" scenario.
Interest rate parity is the basis for exchange - rate pricing. High - interest - rate currencies attract arbitrage capital, and capital inflows drive up the exchange rate. In the 1980s, Paul Volcker's interest - rate hikes pushed the US dollar to a historic peak, and in 2022, Jerome Powell's aggressive tightening also drove the US dollar to a twenty - year high. This logic has withstood the test of time and has never failed.
However, in 2025, the interest - rate differential between the United States and major economies remained at a high level. The yield on the 10 - year US Treasury bond once reached 4.8%, and the federal funds rate was still in the restrictive range. Generally speaking, even if the US dollar did not rise, there was no reason for a significant decline.
But it did decline, and it did so without hesitation.
As if overnight, it lost its sensitivity to "high interest rates." Arbitrage trades were unwound, and capital did not pour in to chase the interest - rate differential; instead, it quietly left the market.
Why?
The Bankruptcy of Old Theories and Models
I think it is first related to the safe - haven property.
Over the past half - century, the US dollar has been the only safe haven during global turmoil. During the European debt crisis, funds flocked to the US dollar. The US dollar shortage at the beginning of the COVID - 19 pandemic is still vivid in people's memory. This safe - haven premium is a direct manifestation of the core competitiveness of the US dollar system.
However, the situation in 2025 was completely different.
After the Trump administration announced "reciprocal tariffs," global trade frictions suddenly escalated, geopolitical tensions intensified, the fire of war reignited in the Middle East, and the situation in Venezuela changed suddenly. Traditionally, each of these events should have pushed up the US dollar.
But the US dollar not only did not strengthen; instead, it accelerated its decline.
Mainly, funds flowed in two directions. One was gold, and the spot price of London gold once exceeded $5,500. The other was RMB assets, and the onshore Chinese market recorded significant inflows of foreign capital.
It was unthinkable in the past that safe - haven funds would no longer default to the US dollar.
From the perspective of the balance of payments, the relationship between the trade deficit and the exchange rate has also changed significantly. A trade deficit means a net outflow of US dollars, which puts pressure on the exchange rate. Conversely, narrowing the deficit helps support the US dollar.
In 2025, the US government tried to suppress imports and improve the trade balance through high tariffs. Logically, this should have provided positive support for the US dollar. However, the actual trend was the opposite. After the tariffs were implemented, the trade deficit did not substantially narrow. The "front - loading of imports" effect even led to an expansion of the short - term deficit. The key is that during the policy cycle of "trying to narrow the deficit," the US dollar depreciated faster than during the period of "allowing the deficit to expand."
That is to say, the market's pricing logic for the US dollar's credit has changed. Investors no longer care whether the United States can improve its external balance through trade policies. Instead, they are concerned about another issue.
All these differences indicate that the underlying assumptions of the traditional exchange - rate analysis framework are starting to fail.
Tools such as interest rate parity, purchasing power parity, and safe - haven property premiums have been effective in the past few decades because they are all rooted in a specific historical condition - the monetary order of the era of industrial capitalism.
In that order, the main driving force of capital flow was profit - seeking. Information was slowly transmitted through the correspondent banking network, and the endorsement of sovereign credit had almost absolute authority.
When analyzing money in Das Kapital, Karl Marx pointed out that the ability of money to act as a general equivalent ultimately depends on the trust of the commodity world in the abstract labor time it represents.
In the era of industrial capital, this trust was relatively stable because it was anchored in the production and exchange of physical goods.
However, digital financial capitalism has changed all this.
Capital still seeks profit, but "survival security" has replaced "maximizing returns" as the primary consideration for sovereign capital and long - term institutional investors. When the US fiscal deficit exceeded 6.4% of GDP, the debt - ceiling negotiations became a political show, and the independence of the Federal Reserve was publicly challenged as never before, the risk - return characteristics of holding US dollar assets underwent a structural shift.
Information is no longer slowly flowing and symmetrically distributed. Every sovereign investor can see in real - time that the US policy uncertainty index has soared to its highest level since 1985, and that the risk compensation implied in the term premium is expanding.
Therefore, the essence of the US dollar's weakness in 2025 cannot be attributed to early expectations of interest - rate cuts or short - term fluctuations in the trade deficit.
These are just triggering factors.
What really happened was the international capital's re - pricing of US dollar assets. It was not about lowering the target valuation but replacing the valuation model itself. The old model had "interest rates" and "growth" as core parameters, while the new model is setting "the stability of the credit anchor" and "the irreplaceability of the technological anchor" as higher - weighted variables.
When there are signs of loosening in both anchors, capital outflows are no longer a periodic portfolio adjustment but a structural asset reallocation.
Pre - empted by Open - Source
Moreover, in the pricing logic of US dollar assets, there is a premium that is rarely publicly discussed but is extremely important. It is the consensus expectation of global capital. Previously, people generally believed that the United States would always be the sole cradle of the next - generation technological revolution.
From Intel and Microsoft in the PC era, to Cisco and Google in the Internet era, to Apple in the mobile Internet era, and then to Nvidia and OpenAI in the early stage of the AI wave, the trajectory of technological iteration has never deviated from the US soil in the past forty years.
Global capital holds US dollars and buys US stocks because they believe that no matter what the next round of disruptive innovation is, the exchange where the dividends will be finally realized must be in New York.
In 2025, DeepSeek's open - source release of the R1 model should be marked separately in the history of the US dollar exchange rate.
Its extremely low cost, performance approaching that of GPT - 4, and complete open - source nature have been repeatedly discussed, but the truly disruptive consequences are still fermenting.
This was not a regular catch - up by the Chinese AI industry, nor was it just another victory of the open - source community over the closed - source commercial model. From the perspective of political economy, this was a break in the history of the evolution of global production tools: for the first time, the most cutting - edge general artificial intelligence was no longer the "private means of production" of Silicon Valley technology giants but was publicly issued globally at an almost zero marginal cost.
When analyzing the production of relative surplus value in Das Kapital, Karl Marx pointed out that the source of excess profit is the continuous higher labor productivity of individual enterprises than the social average level. The premise of capital premium is the exclusive possession of production tools.
When Nvidia's computing clusters and OpenAI's algorithm architectures constitute the most core means of production in the era of artificial intelligence, and when such means of production are concentrated within a radius of dozens of kilometers in the San Francisco Bay Area and Silicon Valley, US technology capital has obtained the institutional power to extract global technology rents.
Global capital's purchase of Nvidia stocks is essentially a subscription for the future dividends of this exclusive possession.
And DeepSeek's open - source release, through a straightforward technological diffusion, deconstructed this exclusivity at the property - rights level.
The consequences are first reflected in the valuation logic of US technology giants.
The core assumption supporting the six - trillion - dollar market value of the "Magnificent Seven" is not how much money they are currently making, but that they have locked in the entry tickets to the next - generation general technology platform. When this entry ticket is no longer a scarce item with a limited supply but an open - source resource freely available to global developers, the "scarcity premium" in the valuation system loses its foothold.
In January 2025, Nvidia's market value evaporated by 600 billion US dollars in a single day. That was the price set by capital after re - evaluating the "degree of privatization of the means of production."
Over the past twenty years, the pattern of global US dollar reserve accumulation has changed. Before 2008, trade - surplus countries mainly allocated the US dollars they earned to US Treasury bonds. After 2008, with the start of the zero - interest - rate era and the long - term bull market in US stocks, the equity assets of US stocks held by foreign investors quickly climbed to a scale of two trillion US dollars, exceeding the holdings of US Treasury bonds for the first time.
This means that the demand structure supporting the US dollar exchange rate has also changed. The allocation of official reserves is stable and has low elasticity, while the equity allocation of the private sector is highly liquid and extremely sensitive to valuation changes.
When there are cracks in the "technology monopoly narrative" of US technology giants, the willingness of this part of global capital to allocate US dollar assets is no longer stable.
It's not that the Federal Reserve cut interest rates too slowly, nor is it that the US economy is in recession. It's that the core story that made global capital willing to accept a lower dividend yield and a higher price - earnings ratio has been weakened in credibility by an open - source release.
This weakening will further erode the long - term fundamentals of the US dollar exchange rate through the transmission of the industrial - chain division of labor.
The traditional "smiling curve" endows the US dollar with an industrial foundation that seems unchallengeable. For example, the United States occupies the high - value - added links of R & D, design, and brand terminals, East and Southeast Asian countries undertake the low - value - added links of manufacturing and assembly, and oil and commodity - producing countries are at the bottom. This division - of - labor order supports the strength of the US dollar. Since the highest - value - added production links are only completed in the United States, the world naturally needs to hold US dollars to purchase these finished products and services.
The disruptive impact of open - source AI is that it has significantly lowered the specialized threshold of the R & D link.
When a hardware startup in Shenzhen can use the same - level open - source models as its Silicon Valley counterparts for chip design, structural optimization, and market analysis, the scarcity premium of the R & D link begins to converge.
The arc of the smiling curve is being flattened.
In this regard, the cyclical exchange - rate fluctuations do not belong to this phenomenon. I think it is more of the starting point of a structural trend. The United States' niche at the top of the industrial chain is no longer unique, and the relative advantage of US dollar assets attached to this niche is no longer permanent.
The underlying logic of global capital selling US dollar assets in 2025 is not a negation of the United States' technological capabilities, but just a disenchantment of the "uniqueness." When the world discovers that the means of production for the most core knowledge tools of the future are no longer monopolized in the land between the Ohio River and the Pacific Ocean, the hidden technology option fee of the US dollar exchange rate, which has never been included in the interest - rate parity model, begins to be written off by the market one by one.
The Dissipation of Digital Currency and Network Externalities
Beyond AI, there is another deepest moat of the US dollar hegemony, an invisible and intangible network that is almost impossible to escape.
The name of this network is SWIFT. It is a financial messaging system, not an actual currency settlement system.
It only transmits instructions and does not process funds.
Its terrifying aspect does not lie in the speed of message transmission, but in the set of correspondent - bank clearing order it has solidified.
In this order, a trade payment from Jakarta to Lagos must pass through New York. A Colombian exporter in the coffee - bean business must ultimately hold US dollar reserve funds in a custodian bank in Manhattan.
This is the institutional inertia accumulated by path dependence.
The settlement instructions for 370 trillion US dollars of global financial transactions are all completed in this topological structure.
Any country that tries to challenge this system initially aims to build a "better SWIFT" - faster messaging, more secure encryption, and fairer governance.
But this idea is wrong from the start.
The digital RMB and the mBridge project have precisely avoided this trap.
It is not competing with SWIFT in terms of who can send telegrams faster, but is changing the way of communication itself.
Traditional cross - border payments rely on nested correspondent - bank accounts. Each layer occupies liquidity, charges fees, and exposes to the judicial jurisdiction of the other country.
While mBridge constructs a peer - to - peer distributed clearing logic: two commercial banks, whether located in Beijing or Dubai, and whether holding RMB or Thai baht, can directly complete delivery versus payment on a shared platform based on smart contracts. This is not an improvement of the correspondent - bank model but a complete abolition of the intermediary status.
When the oil buyers of Saudi Aramco can directly pay the Chinese refineries with digital RMB, and the funds can complete cross - border settlement within a few seconds without passing through the judicial jurisdiction of any third country, the US dollar as a medium of exchange is simply skipped.
SWIFT is still operating, and New York is still clearing, but this transaction has nothing to do with them.
What really makes the US dollar system uneasy is not only the technical parameters of mBridge but also the strategic anxiety exposed by the United States in response to this trend.
The GENIUS Act passed in 2025 was packaged by Western media as a milestone in digital - currency regulation, but few people have dissected the real intention of this act. It is not to stop the development of private stablecoins. On the contrary, it is to incorporate USDT and USDC, which have grown wildly to a scale of hundreds of billions, into the US dollar system, turning them from potential competitors into mercenaries on the US dollar's digital frontier.
This is an extremely delicate defensive stance.
In 1971, when Nixon closed the gold window, the United States was on the offensive - actively tearing up the Bretton Woods rules to gain more freedom in monetary policy.
In 2025, with the GENIUS Act, the United States is on the defensive. It cannot prevent other countries from developing their own central - bank digital currencies, nor can it prevent the penetration of non - US - dollar stablecoins in cross - border payments. So it chooses to bring its own private crypto - dollars under regulatory scrutiny, trying to exchange compliance licenses for market survival.
But the price paid for this incorporation is a subtle downgrade of the US dollar's credit foundation.
Previously, the world used the US dollar, essentially trusting the US government's tax - collection ability, repayment willingness, and the rule - of - law tradition. This is national credit, a public credit accumulated over two centuries and backed by the constitution and budget procedures.
After the GENIUS Act came into effect, the circulation basis of the digital US dollar is the code audit reports of Circle and Coinbase, the reserve - asset custody certificates, and the compliance review of the US Securities and Exchange Commission. The credit subject has been downgraded from the United States of America to private technology companies registered in Delaware.
This is still a high - quality credit asset, but it is no longer an irreplaceable sovereign credit.
If the marginal supply of a currency depends on the compliance costs of private companies and the risk preferences of their boards of directors, the pricing power of this currency cannot be completely retained in the hands of the Federal Reserve. This explains an abnormal monetary phenomenon in 2025 - the expected fluctuation range of the federal funds rate has narrowed, but the intraday volatility of the US dollar