The market turmoil triggered by the AI investment boom has not yet shaken the fundamental support for the US dollar. According to the latest foreign exchange strategy report from Barclays, US economic data remains consistently strong, and the possibility of a sharp downward revision in the AI capital expenditure cycle in the near term is limited, leading the bank to favor a strategy of buying the dollar on dips.
Currently, the dollar risk premium has fallen to about 1%, the lowest level in more than a year. Analysts point out that EUR/USD may briefly rise to around 1.14 due to crowded euro short positioning, but given the continued resilience of US macroeconomic fundamentals, this move is "unlikely to be sustained."
Uncertainty over AI capital expenditure is affecting major currencies in different ways 鈥?the yen benefits from safe-haven demand, while the euro is further weighed down by French fiscal risks.
Starting from the "shipwreck": a mirror of history
Barclays opens its report with a historical anecdote: in 1858, engineer Isambard Kingdom Brunel led the construction of the SS Great Eastern, then the world's largest ocean-going passenger ship. The vessel drained the finances of its backers, changed hands several times, and was ultimately sold for scrap, while the dismantling company itself fell into delays and labor disputes.
Of course, the story of the Great Eastern is not purely a tragedy. The ship was later used to lay the transatlantic telegraph cable between Britain and the United States, giving rise to a technological innovation that would permanently change the world 鈥?the nickname "cable" for the GBP/USD exchange rate originates from this very episode.
The bank uses this as an analogy for the current AI investment wave, noting that "this may be the first time since the late 1990s that a project of such scale and ambition 鈥?a journey into the unknown in terms of commercial use, profitability, and consequences 鈥?is unfolding before our eyes. Moreover, it is evolving in a way that goes beyond anything our generation has perceived as 'risk-taking.'"
AI capital expenditure: where are the risks?
The bank directly identifies the most worrying structural issues in this AI capital expenditure cycle.
First, the authenticity of demand is questionable. Is current demand for computing power being pushed artificially high by projects that cannot be sustained in the medium term?
Second, the transmission risk of cross-financing. Could extended cross-financing among chip foundries, AI software companies, and hyperscale cloud providers trigger a domino effect if a single link fails?
Third, political resistance to data center construction. After the US midterm election results, the market is also asking: could the backlash against data center construction cause delays and drive up costs? And how would that transmit to chipmaker valuations?
However, the report notes that existing data center construction contracts will proceed as planned, and construction progress over the next few quarters is unlikely to suffer a substantial shock. Judging from the latest earnings reports, more ample computing power supply is driving revenue growth 鈥?to some extent, this must be driven by real enterprise demand, rather than solely by large language model training.
Hyperscale cloud providers currently show no signs of slowing their borrowing and investment plans; in fact, they are expected to achieve positive cash flow by the end of next year.
More importantly, Barclays offers a bottom-line macro judgment: "This new technology has arrived and will not leave. There is no way we can return to the state of two years ago."
The US dollar: the AI selloff has not yet had a material impact
Barclays' conclusion is that, at least for now, the AI-related stock selloff has not shaken the dollar's appeal.
The report notes that existing data center construction contracts will proceed as planned, and capital expenditure over the next few quarters is unlikely to be substantially reduced. Hyperscale cloud service providers currently show no signs of slowing their "borrow-to-invest" plans and are expected to achieve positive cash flow by the end of next year. In addition, higher computing power supply is driving revenue growth, which to some extent reflects real enterprise demand rather than being driven merely by large language model training.
From a macro perspective, the growth rate of US private domestic final sales to purchasers is close to 5%, the economy is operating above potential output, and inflation remains sticky near 3%. The bank believes that "the capital expenditure cycle will continue, with few signs of a sharp downward revision over the next year."
Current dollar sentiment indicators remain in slightly positive territory, with no net short signal yet. The bank says the market may see a narrative that "the US is about to slow," pushing EUR/USD briefly to around 1.14, but "given the continued resilience of the US macro backdrop, both currently and prospectively, this move is unlikely to be sustained."
The euro: its appeal as an anti-dollar alternative continues to fade
The euro's appeal as the main "anti-dollar" alternative currency is fading at an accelerating pace.
Barclays points out that the euro's weakness already began to emerge after the Federal Reserve's September rate hike, and the sharp widening of French OAT spreads has further added pressure.
France's fiscal problems are already thorny enough, and the upcoming presidential election and a hung parliament make building consensus an almost impossible task.
On the monetary policy front, after the AI-related stock selloff, market expectations for the European Central Bank's terminal rate have been revised to a more realistic level of about 3%. But the bank argues that the euro's monetary policy dilemma lies in this: given the loss of aggregate demand momentum and the fragility of the French bond market, a further significant rate hike above 3% is unlikely.
The medium-term outlook remains negative, and the only near-term bright spot is that sentiment is still bearish, leaving some room for a technical rebound.
The yen: safe-haven demand provides support
Concerns about a slowdown in AI investment have brought safe-haven inflows into the yen through the widening of equity risk premiums. Falling oil prices are also positive for the yen in principle, although this effect may be more evident in yen crosses against other currencies rather than against the dollar.
The bank provides specific scenario estimates: if the S&P 500 extends its current decline of about 6% to 10% (the average pullback since 2023), USD/JPY could fall by about 1.5% toward 155; if the decline widens to 20%, the yen could appreciate by 4% to 5%.
However, the bank also notes that since the coordinated intervention in the summer, speculative positioning has shifted overall to net long yen, which constrains the pace of further yen appreciation.
Other Asian currencies: diverging effects from tech stock volatility
The turmoil in AI and tech stocks is creating clear divergence in Asian currency markets.
The South Korean won (KRW) is also being dragged down by tech stock volatility, with foreign equity net outflows of about $5.9 billion this month, plus about $2 billion in retail net outflows. Barclays believes USD/KRW faces near-term upside risk, with 1325 as support.
The Indian rupee (INR) is relatively benefiting: the sharp drop in oil prices, lower AI exposure in Indian equity indices, and improved foreign equity inflows (about $2.4 billion in net inflows during the month) provide a favorable environment for the rupee. But Barclays remains cautious about the Reserve Bank of India's (RBI) expanding forward book and declining foreign exchange reserves, and expects USD/INR will not remain below the 200-day moving average of 94.22 for long.
The Malaysian ringgit (MYR) faces dual pressure: as the region's only net energy exporter, falling oil prices are relatively more unfavorable for it; at the same time, concerns about data center investment could also weigh on Malaysia's appeal as an Asian AI and data center hub.
Comments