The economy isn’t broken. It’s hollowing out from the inside — and the thing holding it together is more fragile than people realize.

Ten months ago, a lot of people (including me) were calling for a consumer-led recession. Savings depleted, consumer confidence cratering, wages flat against inflation. The thesis was simple: the average American is tapped out, and when 70% of GDP is consumer spending, that's a death sentence.

That recession didn't happen. But something arguably worse did.

Here's the rundown:

• The K just became an E: It's no longer rich vs. poor. Bank of America's data shows three tiers of consumer are now diverging. High-income spending grew 2.5% YoY in January. Middle-income: 1%. Lower-income: 0.3%. The middle class isn't collapsing — it's slowly detaching from the top and drifting toward the bottom. Economists are now calling this an "E-shaped" economy.
• The top 20% IS the economy: The wealthiest fifth of households hold 72% of total household wealth. Consumer spending outpaced disposable income last year, meaning even the people who are spending aren't doing it from earnings — they're doing it from asset gains. Stocks, home equity, AI-era portfolio growth. Take that away and the spending engine seizes.
• The One Big Beautiful Bill made it worse, not better: The OBBBA tax cuts were supposed to help. And they do — if you're already doing well. TD Economics says the benefits are "skewed toward higher-income households." No-tax-on-tips and the child tax credit help at the margins, but they don't offset the structural divergence. The bill entrenched the K-shape rather than resolving it.
• Consumer confidence is in the basement: Michigan Consumer Sentiment hit 55.5 in March — that's the 2nd percentile of the entire series history. Not 2nd quartile. 2nd percentile. People across every income group, age bracket, and political affiliation reported weaker expectations for their personal finances. This isn't partisan vibes. This is broad-based pessimism.
• Savings are still thin: Personal savings rate ticked up to 4.5% in January, but that's still historically low. 13% of adults have zero emergency savings. Lower-income households burned through their pandemic cushions years ago and never rebuilt them.

Why this matters now: The entire economy is being carried by one channel — asset wealth driving upper-tier spending. And that asset wealth is overwhelmingly driven by AI-era equity valuations.

This is where it gets scary.

The load-bearing wall is AI capex. Here's the chain: Massive AI infrastructure spending → inflated corporate earnings and forward guidance → record equity valuations → wealth effect for top-20% households → sustained consumer spending → GDP growth. Break any link in that chain and the whole structure shudders.

And now, two things are threatening to break it simultaneously:

• Geopolitical shock: The US military conflict with Iran that started February 28 sent energy prices surging. Oil price spikes act as a tax on every layer of the economy. For lower-income households, it's immediate — gas, groceries, utilities. For the asset-rich, it's indirect but potentially more destructive: energy shocks historically trigger equity repricing, risk-off rotations, and a reassessment of forward earnings assumptions.
• AI growth repricing risk: The market has been pricing in a "golden age" of AI-driven productivity. But AI capex is a bet on future returns that haven't materialized yet in most sectors. If geopolitical instability raises the cost of capital, compresses risk appetite, or simply shifts attention from "what AI could do in 3 years" to "what happens to my portfolio this quarter" — the narrative can turn fast. Deloitte is already modeling a downside scenario where AI investment becomes "overdone" and triggers a business spending pullback. Morgan Stanley assigns a 15% probability to a mild recession. The question isn't whether AI is real. It's whether the valuations built on AI assumptions can survive a multi-front stress test.

What kind of downturn are we looking at?

Unlike what I said 10 months ago, this doesn't look like a traditional consumer-led recession. It looks like something new: a wealth-effect recession, where the economy doesn't break because people stop spending — it breaks because the asset prices that enable the spending correct, and there's nothing underneath to catch it.

The bottom two-thirds of the income distribution is already running on fumes. They're not driving GDP growth. If equity markets correct 15-20% on geopolitical risk or an AI narrative reset, the top tier — the actual engine — pulls back. And then you get a demand shock that hits an economy where the middle and bottom have no buffer.

Consumer spending can sustain a K-shape "for some time," as Bank of America puts it. But "for some time" isn't forever, and the risks are now stacking — energy prices, a hot war, sticky inflation, an AI valuation cycle that may be peaking, and a consumer base where 72% of wealth sits with 20% of households.

TL;DR: The recession bears from last year were wrong on timing but right on fragility. The economy didn't break from the bottom up — it hollowed out. Now it's standing on one leg (AI-driven asset wealth) while someone lights the floor on fire (Iran, oil prices, tariff overhang). The question isn't whether the average American is struggling. They are. The question is how long the wealthy can keep carrying GDP before something knocks them off balance too.

What are your thoughts? Is the AI wealth channel sustainable through a geopolitical crisis, or are we watching the setup for a correction in real time?