The AI boom has funnelled global savings into US assets, making the dollar a crowded bet on American exceptionalism. If the trade unravels, capital could flow out of the US and towards overlooked markets such as India. But emerging markets are no longer one trade, and the winners will depend on relative growth and earnings.
The next big test for the US dollar may come not from the Federal Reserve, but from the artificial intelligence trade. The AI boom has drawn foreign capital into US mega-cap stocks, hyperscaler bonds and the infrastructure powering the technology revolution. Those flows have supported the dollar by reinforcing expectations of stronger US growth, higher interest rates and continued demand for American assets.
But what happens if the trade reverses? The same capital that has been flowing into the US could begin moving out, weakening demand for the dollar. And unlike earlier market crises, when investors rushed towards the greenback, this time the selloff could be centred on US assets themselves. For India, which has been overlooked as global investors chased AI-linked opportunities in the US and parts of Asia, the reversal could offer a chance to regain foreign capital.
The argument is no longer confined to market speculation. Most experts believe that the foreign money chasing returns from the AI boom could weaken the dollar's traditional safe-haven role if the trade unwinds. “The dollar will most certainly not benefit in the same way it did in the past global crisis’,” says Ritesh Jain, Founder, Pinetree Macro, a hedge fund based in the United States. The world has too many eggs in one basket
The scale of the capital concentration is central to the dollar-bearish argument. Rob Subbaraman, Nomura's managing director and chief economist for global macro research, says the AI boom has masked a rising US risk premium by drawing an extraordinary share of global savings into American assets. His research shows that the ratio of US net international investment position (NIIP) liabilities to the combined NIIP assets of all net creditor nations has surged to 80%.
In other words, a much larger proportion of the savings accumulated by countries that lend to the rest of the world is now invested in the US. In an interview with Moneycontrol, Subbaraman put it starkly: “The world has too many eggs in the one basket, i.e. the US.” He says that in 2010, around 20% of the overseas investments of net creditor nations were held in the US. Today, that share has seen a fourfold increase, driven substantially by the AI story.
“I do think over time, we're going to start to see that get diversified out more.” The vulnerability is not just the size of foreign holdings, but the nature of those investments. Nomura's report notes that US portfolio liabilities — foreign holdings of equities and debt securities — quadrupled to $37.4 trillion in March 2026 from $9.2 trillion in March 2009. A substantial part of that increase has come from capital gains on AI-driven US equities.
If investors begin to cut their exposure, even a moderate fall in foreign holdings could have a disproportionate impact on the dollar. Subbaraman's analysis argues that declines in US portfolio liabilities could exceed any corresponding decline in US holdings of foreign assets, resulting in a net reduction in the country's external liabilities and dollar depreciation. Whoever had to put money in has put it in
Ritesh Jain makes a similar case, but starts with the sheer scale of the capital drawn into the AI trade. “When AI companies' share prices rise, whether through equity or their ability to raise bonds from overseas investors, that means more demand for the dollar,” he says. He points to SpaceX's $40 billion fundraising, which attracted European investors, as an example. “What will a European do now?” he asks.
Ditto for other countries that have chased the AI trade. Jain argues that Middle Eastern sovereign wealth funds, which have invested heavily in US technology and AI infrastructure, can’t continue deploying capital abroad at the same pace either. These funds have substantial domestic investment and rebuilding commitments, he says. “They were the biggest pool of capital,” he argues, and their priorities could shift towards their own economies.
“As long as the AI trade continues to do well, US will suck away the savings or capital from the rest of the world. You have seen that happen.” Now, it has to reverse, he adds. He argues that the traditional argument that capital will flee back to the US in times of crisis is set to be broken this time. In 2008, when the Global Financial Crisis occurred, the US accounted for about 35% of global GDP and 35% of global market capitalisation.
Today, its share of global GDP is around 30%, but its share of global market capitalisation has risen to 65%. “This means that whoever wanted to invest in the US has already invested,” he says. He also points to the US NIIP. “US NIIP was actually kind of positive in the year 2008. Today, US NIIP is minus 60%,” he says, adding that foreign investors hold $24-25 trillion more in US assets than the US holds in assets overseas.
The pool of overseas savings available to keep financing the US, he argues, is not unlimited. Jain sees another sign of this in the dollar's relatively modest rise despite oil prices climbing to around $115 a barrel given that the US is a net exporter of oil. “The dollar should have been stronger already, but it is not,” Jain says. “When the AI trade unravels, the real weakness will come to the fore,” he adds.
This time, the dollar may not be the automatic refuge Interestingly, while foreign flows into US equities have reached record high, foreign direct investment has lagged. Much of the money is therefore coming from institutional investors who can reverse portfolio allocations when valuations or growth expectations change. The consequences could play out through three channels. A retreat from US equities and corporate bonds would reduce demand for dollars; weaker AI investment would hit US growth; and a slowdown could prompt the Federal Reserve to cut rates, eroding the dollar's yield advantage.
The market has already offered a glimpse of this possibility. Reports suggest that during a recent bout of AI-related volatility, safe-haven flows favoured the Japanese yen and Swiss franc over the dollar. Subbaraman cautions against assuming that the dollar will always benefit from a crisis. A correction could set off a feedback loop in which falling asset prices weaken confidence and growth, prompting further selling.
That does not make a dollar slide inevitable or immediate. A severe funding squeeze could still produce a temporary rush for dollars. But if the AI correction becomes a US-specific reassessment of growth, valuations and capital allocation, the currency could be on the losing side. Emerging markets are not one trade For India, the opportunity lies in the possibility that global investors begin to diversify away from an increasingly crowded US trade.
But a weaker dollar does not mean all emerging markets will benefit equally. At the Rising Bharat Summit, Morgan Stanley India's Ridham Desai highlighted the sharp divergence in the performance of emerging markets. South Korea and Taiwan have benefited from the AI and semiconductor boom, while India has been left behind. Korea's market, he noted, had fallen around 35% from its peak even as foreign investors had sold about $150 billion of equities this year, despite strong AI-linked earnings.
The point is that emerging markets cannot be treated as one homogeneous asset class. Their exposure to technology, sector composition and earnings trajectories vary sharply. A reversal in the AI trade could hurt the markets that benefited most from the boom; it does not follow that the same money will flow indiscriminately into every other emerging market. Subbaraman makes a similar distinction. The AI investment boom, he says, has delivered “disproportionate benefits” across countries.
The US has gained substantially, as have parts of Asia, including Korea, Taiwan, Singapore, Malaysia, China and, to some extent, Japan. Europe and India have benefited much less. “If there is a big correction, those that benefited earlier get hit hardest, but the ones that haven't benefited, don't get hit as hard,” he says. That divergence could create room for a more diversified allocation, says Subbaraman.
Why India could be in the frame India has been at the receiving end of the AI-driven capital shift. Investors who found stronger earnings growth in US technology companies and Asian semiconductor businesses had little reason to increase exposure to a market that was delivering relatively weaker earnings growth. That equation may now be changing. After two years of lacklustre earnings, India's earnings growth is beginning to accelerate, while some of the markets that benefited from the AI boom may face tougher comparisons and a slower pace of growth.
While growth in those companies may still be fare stronger, markets react to expectations, thus lower than expected growth will be negative for AI stocks. Subbaraman believes India's fundamentals remain sound. “India's got decent growth, long-term growth around 7% in our view right now. We feel that monetary and fiscal policies are prudent. It's got a solid financial sector,” he says. “So it's a pretty healthy, fundamentally strong economy.
And as long as that's sustained, I think India will ultimately start attracting more capital inflows.” He also sees an opportunity in the next phase of AI adoption. The first phase has been about building the infrastructure and producing AI — an area dominated by the US and a handful of Asian economies. The next will be about using AI to improve productivity across businesses and economies. “But then how do you use the AI?
Because that's really where you get the productivity gains,” he says. “It's not certain that that's going to be the US. It could be other countries that are much better adapters of using the AI and diffusing it into companies and economies.” That could work in India's favour. The country need not replicate the massive capital expenditure of US hyperscalers to benefit from AI; it could capture gains through adoption across services, manufacturing and other sectors.
For Subbaraman, the policy test is to stay the course. “If India manages to keep going with reforms, and keeps monetary and fiscal policies prudent, and the financial sector healthy, I think over time, capital inflows will be coming back to India.” The dollar's decline, if it comes, will not by itself guarantee a return of foreign money to India. Oil prices, domestic earnings, valuations and the relative attractiveness of other markets will matter.
India could benefit from a reallocation that has little to do with a sudden discovery of its merits — and much to do with investors looking beyond an increasingly crowded trade. And finally, India could find itself on the right side of one of the biggest global portfolio reallocations in years.