Draft for editorial review — approx. 1,500 words
The headline from last week’s GDP report will be repeated everywhere: the U.S. economy grew at 1.5% in the second quarter, down from 2.1% in the first. That framing is accurate and almost beside the point. The real question isn’t whether 1.5% is good or bad in isolation — it’s what’s holding the number up, and whether that support is durable. Pull back the curtain, and Q2 was carried by exactly one engine: the American consumer, spending at a 3.2% annualized clip, up from a nearly stalled 0.5% in Q1. Business investment helped too, but a meaningful share of it is now concentrated in a handful of companies racing to build AI infrastructure, which is a different kind of dependency than it looks like on a GDP table.
That is not a diversified economy. It’s a single-engine plane, and the engine is consumer credit, a generous tax refund season, and a few weeks of cheap gasoline that have already reversed.
Start with government spending, because it complicates the picture more than it gets credit for. Federal spending actually fell 4.1% in Q2, with non-defense spending down nearly 13%. On its face, that’s fiscal restraint. But it sits awkwardly next to a resumed and escalating war with Iran, one CENTCOM is reportedly planning to extend for weeks, not days. Wars aren’t cheap, and they aren’t deflationary. If Q2’s spending decline was a lull before wartime outlays ramp back up, that’s a very different trajectory than “the government is tightening its belt.” I’d bet on the former. Governments rarely cut spending mid-war; they borrow and spend more, and the bill eventually shows up in Treasury issuance and the deficit conversation both parties like to avoid until they can’t.
Then there’s the stock market, doing something odd relative to the real economy underneath it. The Magnificent Seven now make up roughly a third of the S&P 500’s value — the highest concentration in the index’s modern history — and their combined capex could top $750 billion this year. That’s large enough to move national business-investment figures, meaning the stock market and GDP are no longer just correlated; they’re mechanically linked through the same handful of balance sheets. When Amazon spends $200 billion on AI infrastructure without the free cash flow to fully cover it, that capex still counts as investment in the GDP math — increasingly funded by debt and share issuance rather than operating profit. That’s a subtle but important crack in the “AI is different because it’s cash-funded, not debt-funded” argument bulls have leaned on for two years. If that argument is starting to give even at the edges, the capex supercycle propping up both the market and a chunk of GDP growth is more fragile than the headline confidence suggests.
Here’s the part that should worry anyone doing real financial planning rather than reading headlines: none of this looks urgent in the data everyone checks first. Unemployment ticked down to 4.2% in June. On its face, good news. It isn’t. It fell because people left the labor force, not because more found jobs. Payrolls grew by just 57,000, badly missing expectations, and the two prior months were revised down a combined 74,000 — a net negative revision, meaning the labor market has quietly been weaker than we thought for three straight months. Participation is at its lowest since March 2021. That’s not strength hidden by a flattering headline number; it’s a labor market where exits are outpacing entrances.
Layer inflation and rates on top, and the timing gets uncomfortable. Core inflation, by both CPI and the Fed’s preferred PCE gauge, sits well above the 2% target — 2.6% and 2.8% respectively. Normally that alone wouldn’t alarm anyone; it’s been elevated for a while and the Fed has learned to live with a slow grind lower. What’s changed is oil. As recently as mid-July, futures markets priced a reasonable chance of a Fed rate cut later this year. Within two weeks, as fighting with Iran intensified again, that flipped to pricing in a roughly 75–80% chance of a hike in September. That’s an extraordinarily fast repricing, and it means the Fed’s next move is now hostage to geopolitics rather than domestic data. If oil resumes climbing toward the $100–115 range seen at the height of the conflict, we could see the Fed forced to raise rates into a labor market already showing cracks — stagflation-adjacent conditions, not a textbook soft landing.
One more variable deserves attention: tariffs. ISM’s own manufacturing survey commentary in June flagged tariffs as a top negative theme, right alongside the Iran war. Tariffs push input costs higher through a different channel than oil, hitting many of the same manufacturers simultaneously. If tariff escalation and oil-driven inflation collide in Q3, the Fed’s September decision gets harder still — responding to two distinct inflationary pressures with one blunt tool. That’s the kind of setup where a single hike doesn’t fix anything, and a series starts to look necessary, which is exactly what turns a soft patch in hiring into something worse.
So what does that mean for Q3? I don’t believe the Atlanta Fed’s initial 5.0% GDPNow estimate deserves the attention it’s already getting. Early-quarter nowcasts run on almost no real data, and this one is very likely a mechanical artifact of Q2’s trade and inventory swings reversing themselves, not a signal that growth is about to triple. Treat it as noise until at least the September update.
My own expectation — a forecast, not a guarantee — is that Q3 GDP growth lands somewhere between 1.5% and 2.5%, closer to the low end if oil stays elevated and hiring keeps cooling, closer to the high end if there’s genuine Iran de-escalation and gas prices retreat. The jobs reports due August 7 and September 4 will tell us more than any single indicator whether Q2’s “resilient consumer” story can repeat, because that spending was partly financed by one-time tax refunds and a gas-price dip that has already reversed. Consumers can’t spend money they don’t have indefinitely, and a savings rate near 4%, down from over 6% two years ago, means the cushion is thinner than it was. Credit card balances sit at record highs. That’s not a crisis yet — delinquency rates have actually improved on the most reliable measures — but it does mean less room to absorb another shock, whether that’s a gas spike, a soft labor market, or both at once.
If I had to bet on the single most likely Q3 story, it’s this: growth holds up on paper because AI-related capex and a rebound in trade math flatter the number, while the labor market keeps softening underneath it, oil stays a wildcard tied entirely to how the Iran conflict evolves, and the Fed chooses between hiking into a weakening jobs market or holding steady and hoping oil cooperates. Neither is comfortable, and neither is the kind of environment where “the economy grew 1.5%, or maybe 5%” tells y