According to Goldman Sachs strategist Ioannis Blekos, S&P 500 earnings-per-share growth is actually running closer to 26% year-on-year, once you strip out the "other income" mega-cap tech books from the appreciation of its own equity stakes. The headline figure including those gains is 45%.

In other words, this "record" earnings season we keep hearing about only hits 45% because Nvidia and its peers are booking the paper gains on their own investment bets as profit.

Excluding that accounting effect, Big Tech marking its own venture stakes to market, the number collapses by nearly half, to 26%.

The ISM Manufacturing PMI came in at 55.6 for July, against the 54.0 expectations. It is the seventh straight month the index has sat in expansion territory, and is now at its peak in 4 years, a run that marks a sharp break from the pattern of 2022 to 2025.

Why manufacturing strength shows up in the markets? Manufacturing strength tends to show up in financial markets before it spreads through to the rest of the economy. As orders pick up at factories, the effect ripples outward to component suppliers, logistics firms, equipment providers, and the broader web of businesses that service them. Higher output lifts revenue, supports hiring, and widens profit margins across that chain.

That cash doesn't just sit on balance sheets. Companies plow it back into capacity, add workers, and, over time, some of it flows into financial markets. It's why a sustained manufacturing upswing has historically coincided with a favourable environment for risk assets.

The trend isn't confined to the US. Japan posted its own strong reading Friday, with manufacturing PMI climbing to 54.5, a seventh consecutive month of expansion there too, and the fastest pace of factory output growth since 2014.

With two of the world's largest manufacturing economies accelerating at the same time, the global industrial cycle looks to be gaining momentum.

This chart shows capital spending on AI is on pace to become the largest investment surge in history, comparable with various previous economic booms such as Britain’s 19^th century railway mania and America’s dotcom bubble.

The scale of the bet yet doesn’t guarantee that the returns will follow.

The five biggest hyperscalers have now committed more than $2.6tn to data centres, chips, and power infrastructure. Alphabet alone owes $811bn, the bulk of it hidden in footnotes.

That money is owed whether or not AI demand shows up as promised. And the market is already starting to price in the risk: the cost of insuring these companies' debt has begun climbing. Credit markets tend to smell trouble before anyone else does.

If the spending outpaces the revenue it's supposed to generate, the first sign of stress won't come from earnings calls, but will appear in the bond prices of the world's biggest companies. 

That’s more than $15bn a day.

“There are two ways to enslave a country. One is by the sword. The other is by debt.” – John Adams

Monthly returns since 1990:

August: -0.49%

September: -0.72%

That being said, buying the dip has paid off for decades, thanks to a strong Q4 rebound on average.

China imported more gold in June than what the world's ten biggest gold-mining countries combined can produce in a month. Spread across a full year, that pace would amount to roughly 60% of all the gold mined on Earth.

This isn't the first time Beijing has moved this aggressively to build up reserves. Right after the Iran conflict began, China bought crude oil at record volumes, taking advantage of the uncertainty. Now gold is getting the same treatment.

Is this about moving away from the dollar? Getting ready for a more divided global financial system? Or just building up reserves more broadly? Whatever the reason, the size of these purchases is hard to ignore.

The question isn't whether China is stockpiling anymore, but what it is stockpiling for.