The Two-Speed World: How AI and Oil Decided Who Grew This Quarter

Analysis

Key Facts

—China. The official factory gauge rose to 50.3 in June, but high-technology manufacturing stood far higher at 53.5, while consumer goods lagged.

—Factory prices. China’s factory-gate prices fell for the first time in six months, a sign of deflation where supply outruns demand at home.

—Canada. April output grew 0.5%, but mining, oil and gas alone surged 2.9%, with oil sands up 6.6% on a rebound from maintenance.

—The next month. Canada’s early estimate for May showed growth slowing to just 0.1% as the resource surge faded.

—The Philippines. An economy that imports almost all its oil approached a stated end-of-June limit on part of its fuel supply.

—The line. Growth this quarter ran on two engines, advanced chips and energy, with ordinary demand flat or falling almost everywhere.

Read the world’s economic figures one country at a time and you get four separate headlines; read them together and you get the two-speed world, an economy split between a narrow set of winners and a broad lagging remainder.

At the close of the quarter, four sets of data landed almost together, and each carried the same hidden shape. Growth, where it existed, came from one of just two engines.

The first engine is the advanced electronics that feed the global build-out of artificial intelligence. The second is energy and the raw materials pulled from the ground.

Almost everything else, the shops and the housing and the ordinary demand of ordinary households, was flat or falling. The quarter-end numbers drew that dividing line more sharply than any single economy could show on its own.

China: a factory rebound with a narrow base

Start with China, because its data is the cleanest illustration. The official survey of factory managers edged up in June, returning to modest growth after a flat month.

Look inside the number and the split is stark. The measure for high-technology manufacturing stood far above the rest of the factory sector, lifted by demand for the chips and equipment that the world’s data centers consume.

The parts of the economy that serve Chinese households told the opposite story. The property sector remained under pressure, consumer-goods production lagged, and factory-gate prices fell for the first time in six months, a sign of the deflation that stalks an economy where supply outruns demand at home.

One economist at Bank of America put it bluntly, saying the hope of rebalancing China toward its own consumers had been dashed, with strong exports masking weak demand at home. The country grew, in other words, on what it sold to the AI boom abroad, not on what its own people bought.

The gap between the two speeds was visible inside a single survey. The high-technology gauge sat several points clear of the overall factory reading, while the sub-index for consumer goods barely cleared the line between growth and contraction.

Beijing has set itself a growth target of between four and a half and five percent for the year. Meeting it on the strength of chip exports alone, with prices at the factory gate now falling, is a narrower and more fragile path than the headline suggests.

Canada: a bounce built on oil

Canada’s figure looked, at first glance, like the opposite of a warning. Its economy grew half a percent in April, its strongest month since the previous July, and enough to end weeks of recession talk.

But the composition matters more than the headline. The single largest contribution came from mining, oil and gas, which surged nearly three percent as oil-sands production rebounded from maintenance shutdowns.

The country’s statistics agency noted that fourteen of twenty sectors grew, so the month was not a one-industry story alone. Yet the muscle behind the number was unmistakably the resource sector, swollen further by the higher oil prices that the conflict in the Middle East had pushed up.

The tell came in the same release. The agency’s early estimate for the following month showed growth all but stalling, slowing to a tenth of a percent, as the resource surge faded and the broader economy showed how little momentum it had underneath.

That early estimate is preliminary and will be revised, so it should be read as a direction rather than a fact. But the direction is the point: a single strong month, carried by oil, followed immediately by a near-halt everywhere else.

Canada’s central bank had held its main interest rate steady in June and faces its next decision in the middle of July. One good month built on a resource rebound is exactly the kind of figure a cautious central banker discounts.

The United States and the Philippines: two ends of one line

The United States sits at the winning end of the divide, but in a lopsided way. Its expansion has leaned heavily on a narrow boom in spending on artificial intelligence, the data centers and chips and infrastructure that a handful of large companies are racing to build.

Around that engine, the broader American economy has been softer, with housing weak and ordinary demand subdued. It is the same shape as China’s, a powerful narrow driver sitting above a flatter landscape, only with the country on the other side of the AI trade.

The Philippines shows what it means to sit at the losing end. It imports almost all of the oil it burns, and it approached a stated end-of-June limit on part of its fuel supply with the vulnerability of a country that cannot produce its own energy.

Put the four together and the line is unmistakable. To grow this quarter you needed to be selling into the AI build-out or pulling energy and metals from the ground, while to be a buyer of imported fuel and a maker of ordinary goods was to be on the slow side of the world.

The single force behind both engines

It is worth asking why these two engines, and not others, are the ones pulling the world’s growth. The answer is that both are being driven by the same underlying event.

The build-out of artificial intelligence is an enormous act of construction. It requires vast numbers of specialized chips, and it requires vast amounts of electricity to run the buildings that house them, which ties the technology engine directly to the energy one.

A data center is, in a sense, a machine for turning power into computation. The same wave that lifts demand for advanced processors lifts demand for the electricity and the fuel and the metals that feed and wire and cool them.

That is why the two winning sectors are not really separate. They are two ends of one supply chain, and the countries that grew this quarter were the ones that sold into some part of it, whether the finished chip at the top or the barrel of oil and the tonne of copper at the bottom.

Why Latin America sits on the fast side of the two-speed world

For Latin America this is not a distant pattern but a direct description of the region’s position. Its economies are, overwhelmingly, producers of exactly the energy and raw materials that fall on the winning side of the divide.

The copper of Chile and Perú, the oil of Brazil, Guyana and Colombia, the metals that go into batteries and grids, all of it sits on the resource engine that carried Canada’s April. When the world pays up for commodities, the region’s export earnings rise with them, as they did on the same oil-price surge that lifted Canadian output.

That is the comfortable read. The uncomfortable one is that the other engine, the advanced-chip and AI economy, is one the region almost entirely does not build, and being a supplier of raw inputs to a boom is a very different thing from owning the boom itself.

The case that the divide will not last

The strongest counter-argument says the two-speed world is a snapshot, not a structure, and that reading it as permanent is a mistake. Every technology revolution begins narrow.

The railways, electricity and the internet all started as concentrated booms that enriched a handful of suppliers before the gains spread outward into the wider economy. On this view the AI engine is at the beginning of that arc, not the end.

The spending looks concentrated now because the tools are still being built; once they are in use across medicine, logistics, services and manufacturing, the growth they produce will broaden, and the sharp line between winners and laggards will blur. There is real force in this.

If it is right, the resource producers of Latin America are not merely feeding someone else’s boom but supplying the physical foundation, the metals and energy, of a wave that will eventually raise growth widely. And yet the same history offers a warning.

Diffusion took decades, not quarters, and plenty of economies that supplied the raw inputs of an earlier boom never captured much of the value that flowed to those who owned the technology. Whether this cycle broadens quickly enough to matter, or leaves the slow side waiting a generation, is the question the June figures pose and cannot answer.

Frequently asked questions

What is the two-speed world?

It is the split, visible across the quarter-end data, between a narrow set of winners and a broad lagging remainder. Growth ran on advanced chips and on energy, while ordinary demand stayed flat or fell.

Why did China’s factory number rise?

The gauge rose almost entirely on demand for high-technology goods tied to the global build-out of artificial intelligence. Its property sector and consumer-goods production stayed weak, and factory-gate prices fell.

Where does Latin America fall in the two-speed world?

The region sits on the winning, resource side, as a major producer of the energy and metals the boom consumes. It does not, however, build the advanced chips at the top of the same supply chain.

Could the divide close over time?

It might, if the artificial-intelligence boom broadens into economy-wide gains as earlier technologies eventually did. The warning is that such diffusion has historically taken decades, not quarters.