The tone and tenor in Washington, DC on the sidelines of the recent IMF-World Bank meetings revealed how divided and uncertain the policy, investor and analyst community remains. One would have presumed economic uncertainty would have peaked during the spring meetings in April, a few weeks after “Liberation Day”. The uncertainty then was deep, but narrow, concentrated on the scale and scope of the Trump tariffs. Now it has broadened ominously. Is the US economy on the verge of a recession or has it bucked one? Are the Trump tariffs even legal? Will the breathless AI build-out become progressively unviable? Are we playing with fiscal fire globally? Is the cherished independence of some central banks under existential threat?
Start with US economic performance. The Atlanta Fed Nowcaster suggests US growth last quarter was tracking an eye-popping 3.9 per cent. Far from a tariff-induced recession, US growth is accelerating! Tempting as it may be, it’s too early to claim victory. Below strong headline growth lie caveats and fissures. First, surging AI investment has been responsible for much of recent growth. Absent that, US growth would have averaged just 1 per cent in the first half of 2025. Second, the US labour market has come to a virtual stall, which is typically a precursor to a recession. Third, US inflation pressures continue to build as tariffs are slowly passed on to consumers, which, alongside slowing job growth, will put more pressure on households.
But, given the conflicting signals, which is the signal and which is the noise? Is the labour market about to break? Or are softer jobs numbers simply reflecting a much lower “breakeven rate” because of the anti-immigration efforts by the administration? Will the AI boom offset the tariff pressures and bail the economy out? And if so, will policymakers draw the wrong conclusion that tariffs were not harmful, prompting more to be rolled out?
Meanwhile, AI mania continues to grip the world. Global investments in “data centres” have jumped from $400 billion in 2024 to an estimated $600 billion in 2025 and are projected to rise to $3-4 trillion a year by 2030. It’s understandable why mania will drive more mania. Some, like Nouriel Roubini, believe AI will result in exponential productivity growth, which, if true, will result in a winner-takes-all outcome. So, the first-mover advantage becomes crucial. Further, as the Richmond Fed notes, current AI investment — when compared to the telecom boom of the 1990s — suggests it’s early days yet. We may only be in the second inning of a nine-inning baseball game. Recognising this, the frenzy in the stock market is giddying. Just 30 AI-related stocks now account for 44 per cent of the S&P’s market capitalisation.
But is the exuberance getting irrational? Eventually, AI investments of the kind envisaged will necessitate meaningful leverage. Will the use-cases be enough to justify these investment levels? Will the use-cases be sufficiently monetisable? And then the elephant-in-the-room question. To what extent will AI be labour-substituting versus labour-augmenting? If it’s the former, will the distribution of income become even more skewed in favour of capital at the expense of labour? What will this imply for future employment, consumption and demand? Are we setting ourselves up for the classic fallacy of composition that Keynes warned of? Is the recent wave of AI-related layoffs just the canary in the coalmine? Ominous uncertainties abound.
The case of Taiwan is instructive. Taiwan has been at the epicentre of the AI boom, exporting servers, GPUs and high-end chips that populate US data centres. Taiwan’s exports to the US are growing at a sizzling 50 per cent pace, pushing GDP growth up to 7 per cent. Yet, because the AI buildout is so capital intensive, there has been no positive spillover to jobs and consumption. Even as AI exports have boomed, consumption growth is averaging less than 1 per cent and consumer confidence is falling! So much so that Taiwan’s government had to roll out a 2 per cent of GDP fiscal package in September in spite of the AI mania. A sign of things to come globally?
If these contradictions were to slowly pervade the global economy, where will the fiscal space come from? Unsustainable fiscal deficits in advanced economies are proving to be a key source of global instability. G7 public debt is already at 125 per cent of GDP and expected to rise to 140 per cent by 2030. The US fiscal deficit is running at almost 8 per cent of GDP for an economy above its pre-pandemic path! Exorbitant procyclicality, anyone? Japan’s public debt is 230 per cent of GDP but that is unlikely to stop the new PM from announcing another fiscal package. Bond markets are growing increasingly nervous, but the politics everywhere seems incapable of orchestrating much-need fiscal adjustment. The gamble with fiscal fire is well and truly on.
Finally, the US Supreme Court has just begun to hear whether President Donald Trump’s invoking of the IEEPA (International Emergency Economic Powers Act) to impose tariffs — both reciprocal and fentanyl — is legal. If it is deemed illegal, the US treasury will have to refund Rs $100 billion of tariff revenues, putting more pressure on the US fiscal. The administration is likely to respond by invoking other statues of the law (sections 122, 232, 301) to impose country-specific and sectoral tariffs. But all this will simply compound confusion and uncertainty. And what happens to the legality of trade agreements based on IEEPA tariffs?
Five years after the pandemic and a year into Trump’s presidency, underlying macro fissures have only widened and uncertainties have only grown. US tariffs are the highest since the 1930s and an inexorable descent towards deglobalisation is on. Emerging markets are still grappling with pandemic-induced scarring and now confront a China Shock 2.0 — an avalanche of cheap Chinese imports. The promise of AI (on productivity) is matched only by its peril (on unemployment) at a time when countries lack the fiscal space to backstop labour markets.
The global community that descends on Washington remains divided and uncertain on both the prognosis and the policy response. But equity markets will have none of this. The AI frenzy is propelling markets to ever-new highs. The pressures and problems of the present are being ignored for the (yet untested) promise of the future. Markets may be priced to perfection. But reality rarely follows suit.
The writer is head of Asia Economics at J P Morgan
