US Recession Warning Builds as Markets Rotate
At 14:49 UTC on 2025-09-12, the tape showed a market that is not breaking but is not agreeing with itself either. NVDA printed 218.29, down 0.03% against the previous close and sitting in the bottom 4% of its daily range. AAPL traded at 332.27, up 1.75%, mid-range at 60%. MSFT added 0.65% to 495.63 and TSLA 0.52% to 365.44, both parked near the middle of their ranges, at 48% and 54%. BTC-USD at 77,469.45 was up 0.35% and pressed the top of its range, at 100% of it. Gold at 4,408.90 was almost exactly flat, up 0.04%, mid-range at 68%. The DAX at 25,568.56 gained 0.82% near its highs, 93% of range, while KWEB at 24.60 rose 0.65% but sat in the bottom 4% of its range. That is rotation, not conviction. Money is moving between corners of the board, with nothing collapsing and nothing confirming. The index trend is drifting lower, the tilt is toward safe assets, and the market is waiting on policy.
What Recession Actually Means
The term “recession” is not a synonym for “a bad quarter”. The technical shorthand — two consecutive quarters of falling GDP — is a rough marker. The distinction matters, because an economy can sit far from a technical recession and still post the highest number of new bankruptcy filings since the pandemic. The bankruptcy data illustrate that point. Tuomas Malinen, a Finnish economist specializing in geopolitics and financial crises, argued in a Substack post this week that the US is precariously close to another downturn, despite booming stocks and strong surface-level growth. [3]. “But we need to acknowledge that the bottom can fall beneath the US economy, practically, in any minute,” he warned. [3]
The Break Scenario
The scenario is not a slow, visible slide — it is a break arriving through credit and the AI trade at once. That is why Malinen reaches for the dot-com comparison. “If the above holds, it would indicate that the US economy is very close to the onset of a recession, which can start abruptly from the collapse of the AI-trade,” he wrote. [3]
Three Barometers
The machinery runs through three barometers. First, US bankruptcy filings. In the 12 months leading up to June 2025, the US recorded more than 600,000 new filings, up 12% from the 12 months leading up to June 2024, according to the US Courts office. [1] It is still well below the peaks after the Great Financial Crisis and the dot-com crash. [1] Second, bond yields — and here the construction matters. Malinen’s private sector yield curve is the spread between Baa-rated corporate bonds with maturities of at least 20 years and the bank prime rate, the rate banks charge their most creditworthy institutional clients. . That curve is close to uninverting, meaning corporate yields are close to surpassing the prime rate. Two mechanisms can push it there: investors demanding more compensation for corporate risk, or rising rate expectations, which hurt risk-asset prices. He reads the curve as signaling an imminent onset of US recession, because it turned positive ahead of the COVID-19 recession, the Great Financial Crisis and the recession of the early 2000s. Third, manufacturing new orders. The ISM manufacturing New Orders Index rose to 56.7 in July 2025, a seventh straight month of expansion. [2] It is the only one of the three flashing positive, and Malinen credits the data center boom.
Why the Warning Matters
The warning deserves attention because the two pictures do not reconcile. Malinen expects the AI boom to break suddenly, even as stocks boom and surface-level growth holds. Both can be true: he may be early, and the data may still be deteriorating.
Winners, Losers, and Deciders
Malinen’s scenario points to a small corner of the economy thriving, but warns that success could break suddenly. The losers are broader. Firms already filing have lost the argument. Companies that refinance into a curve that is uninverting pay more to survive. Long-duration Baa holders are exposed if the spread crosses and stays crossed. Anything priced off the AI narrative is exposed to the narrative breaking. Three sets of deciders matter. Banks set the prime rate anchoring one side of the spread. Credit investors set the other side by the yield they demand. Central banks set the expectations beneath both.
What to Watch Next
First, the crossover itself: the private sector yield curve turning positive. Second, the next US Courts filing count, to see whether the 12% annual increase continues. Third, the ISM New Orders Index, to see whether the data center tailwind survives a quarter of tighter credit. Fourth, breadth — whether the rotation continues, with KWEB pinned near its range lows while the DAX and bitcoin sit near their highs. The view carries uncertainty, and the data will decide.
**### What Just Became Obsolete. The assumption that the post-COVID bankruptcy low was a durable floor is obsolete. Treating “the yield curve” as one Treasury signal while ignoring a private sector curve that turned ahead of three downturns is obsolete. The belief that a data center boom is an economy-wide boom is obsolete — one index out of three is not a recovery. The phrase “far from a technical recession” used as a reason to stop looking is obsolete. And dismissing a bearish call because its author has been bearish before is obsolete — that is a reason to weigh it carefully, not to skip the data.
This is educational analysis, not investment advice.
Sources
1. US Courts
2. ISM
3. Substack