Talking Points
April 24, 2026 | 11:58
AI Wins Another War
Economics
Economics
Is history rhyming again? A year ago, markets and analysts were squarely focused on the roiling trade war; equities were on the defensive until late April; and, almost everyone was marking down their economic forecasts. But something funny happened on the way to a downturn—markets promptly reversed course, rallying strongly through the second half of the year, and the global economy ended up doing just fine with 3.4% growth for all of 2025. As it turned out, a tidal wave of spending on AI infrastructure saved the day, heavily supporting U.S. GDP growth, driving tech stocks higher, and compensating for any bite from the trade war. It also helped that almost all countries quickly succumbed to U.S. tariff demands, and chose not to retaliate, avoiding a full-fledged trade war. |
Fast forward to this year. Markets and analysts have been squarely focused on the war with Iran over the past eight weeks. Equities initially corrected in March, yields backed up, and economists boosted inflation projections and were holding the axe over their growth forecasts for 2026. Yet, even with the conflict unresolved, oil prices hovering near $100 (and up more than 50% from pre-war levels), and inflation already having taken a big step up in many economies, markets have largely shaken off this war, too. And we can yet again largely point to the relentless wave of spending on AI’s backbone as a primary factor behind the rebound. Tech stocks have led the snap-back; after correcting by 13% in a matter of weeks, the Nasdaq has since surged 19% from the late-March lows to a fresh record high this week, to stand a towering 60% above the lows hit just over a year ago. It’s not just the equity market that’s getting a lift from AI spending, as it’s also quite evident in the economic data. And, it’s certainly not confined to the U.S, as a variety of Asian economies have reported an absolute surge in exports over the first quarter of 2026. Japan was the latest, revealing that sales abroad managed to gather steam in March, rising 11.7% y/y. But that was a modest gain compared to some of the emerging economies. In US$ terms, China’s exports popped 14.7% y/y in Q1, while Thailand and Vietnam were up 18-19%. Even more impressively, Singapore was up 35% y/y in Q1, South Korea jumped 38% and Taiwan’s exports soared 49% y/y. The common factor behind these astonishing increases was tech gear, and often led by chips. The Economist reports that China’s memory chip exports were up 174% y/y in Q1, and at $46 billion, now represent the single biggest export item for that economy. We have previously noted that, since the war erupted, there has been a deep split between manufacturing activity (solid) and services (softening) in many major economies (Chart 1). That split fits perfectly well with the point that factories are still running strong to meet all the AI demand, while services have been chilled by consumer caution amid higher fuel prices. That theme seems to be continuing deep into April. The preliminary PMIs for this month showed consistent strength across the board in manufacturing, even as services largely stumbled. For example, the Euro Area’s factory PMI rose to a 47-month high of 52.2 while services slumped to just 47.4, a 62-month low, while Japan’s pair of PMIs was 54.9 and 51.2. The U.S. and U.K. ran a tad against trend with small gains in services, but both also reported even stronger advances in factory activity (to 54.0 in the U.S. and 53.6 in the U.K.). Looking through all of those figures, the main point is that factory activity remained quite robust in April, even as consumer confidence and, in many cases, business sentiment took a turn for the worse. We would point to rollicking demand for everything related to the AI build-out as a primary factor for the resiliency of manufacturing activity globally. |
Notably, consumer spending appears to be holding up remarkably well in many major economies, even with the sag in confidence. At the top of the list, U.S. retail sales fired up 1.7% in March, even as gasoline prices were soaring at a record monthly pace. Yes, the headline figure was juiced by the jump at the pumps. But, all the commentary was around how consumers would be forced to divert other spending to pay for fuel. Yet, sales outside of gas stations were up 0.6% m/m in March, hardly a sign of consumers turning tail. U.K. retail sales excluding fuel also topped downbeat expectations with a modest 0.2% m/m gain last month. The flash reading for Canada also points to a 0.6% increase in March sales, although that may just offset price increases. Still, the early read in most economies is that consumers are hanging in there so far. |
|
Looking ahead, the key question for policymakers is whether the conflict can be resolved in time to avert another big step up in oil prices and keep consumers on track. While the current stalemate in the Strait is not overly rattling oil, many analysts have noted that true physical shortages are poised to emerge very soon if the crude doesn’t begin to flow from the Gulf. As much as the equity market wants to move on from the conflict, a renewed spike in oil beckons if the U.S. and Iran can’t soon reach some kind of arrangement on the Strait. And even the AI boom may not be able to override another sharp run-up in energy costs. The AI boom even featured heavily in Kevin Warsh’s sometimes fiery testimony before the Senate Banking Committee this week. It’s Mr. Warsh’s view that AI can boost productivity heavily, allowing the U.S. economy to grow rapidly without sparking inflation. In his view, that opens the door to lower short-term rates (coupled with an eventual reduction in the Fed’s balance sheet), citing the example of the tech boom in the late 1990s. However, we wouldn’t be the first to point out that a productivity boom could allow an economy to grow faster with less inflation pressure, but that’s not really a case for lower rates. If anything, strong real growth lifts the neutral interest rate, which over time could point to higher short-term rates. Indeed, while the Fed did hold off in raising rates in the late 1990s initially, the tech boom eventually led to a series of hikes in 1999/2000 to a peak of 6.5%. And that was at a time when headline inflation was only a bit higher than today, peaking at 3.5% in 2000Q3. The path to Warsh becoming Chair appears to have been cleared, as the DoJ dropped its probe (finally) into Jay Powell on Friday. With this unsavoury episode presumably put to bed, next week’s FOMC meeting may well be the last for Powell, at least as Chair. There isn’t much debate around what the Fed will do this time: markets are priced for only a small chance of even one cut through the rest of the year. Warsh may lean to lowering rates, but the other 11 members of the FOMC may have other thoughts, as does the inflation-pumping level of oil prices, not to mention the ongoing strength in AI-related spending. The Bank of Canada is similarly expected to stay firmly on hold next week. The big debate is whether Governor Macklem will lean hawkishly in his remarks, with the market still pricing in a hike later this year. This week brought a couple of big strikes against the hawks. First, CPI was nicely below expectations in March at 2.4% y/y, and all measures of core came in mild. That’s three consecutive months of decent CPIs, with prices ex food, energy and taxes up at a mere 0.5% annual pace over that time. Second, the looming USMCA review surged back onto the front page, with some heated rhetoric on both sides, pointing to some turbulence ahead. Canada’s lead negotiator counselled calm, but businesses may heed that call to hold their nerve by also holding their spending. With trade uncertainty still dragging heavily on Canada’s economy, we continue to believe that the best course of action for the Bank is zero action. |