Remember when everyone was convinced the Fed would cut rates forever? Yeah, that was last week.
Last Friday, something wild happened: futures markets briefly flashed a 50% probability of a rate *hike* by year-end. For context, that’s like your gym buddy suddenly announcing he’s training for a marathon after years of skipping leg day. It’s a shocking reversal.
The odds have cooled to about 10% now, but the shift itself is the real story. Here’s what spooked the market:
The Triple Threat
First, crude oil hit $110 a barrel as the Iran situation drags on. Brent’s sitting at $114. That’s not great when your economy doesn’t need higher energy costs.
Second, import prices jumped 1.3% in February—the biggest monthly spike since March 2022. Exports rose 1.5%, the largest gain since May 2022. Translation: inflation is knocking on the door.
Third, the OECD just revised U.S. inflation forecasts *up* to 4.2% for the year. The Fed’s own projection? 2.7%. Someone’s math isn’t adding up.
The Stagflation Squeeze
Here’s where it gets spicy: recession odds are climbing simultaneously. Moody’s puts the probability of a downturn in the next 12 months at 50%. Goldman Sachs says 30%. Other firms are calling 40% or higher.
So we’ve got rising inflation *and* rising recession risk at the same time. That’s stagflation—the economic equivalent of being stuck in traffic with a flat tire. The Fed’s in an impossible position: cut rates to save the economy, and inflation runs wild. Raise rates to fight inflation, and you trigger a recession.
The FOMC meets April 28-29. A hike then is unlikely (2.6% odds), but the *expectation shift* matters for every asset class.
The AI Monetization Problem Nobody Wants to Admit
Meanwhile, OpenAI just killed Sora, its video-generation model, after only six months. Downloads had plunged 75% from their November peak. The company was burning $15 million *per day* to run it.
From OpenAI’s own head of Sora: “The economics are currently completely unsustainable.”
This isn’t just one company pivoting. Walmart ditched its ChatGPT shopping integration because it didn’t boost sales. Nvidia’s reconsidering chip commitments to OpenAI. Oracle scrapped a data center expansion with them.
The pattern’s clear: building AI is working. Monetizing it? Not so much. The infrastructure layer (chips, data centers) is generating real revenue. The consumer-facing application layer? Still unproven.
The Private Credit Time Bomb
Here’s where the Fed story connects to everything else: private credit funds are sitting on billions in loans made when rates were near zero. Inflows into these funds crashed from $1.8 billion in early 2025 to $1.1 billion in early 2026. Investors are spooked about exposure to SaaS companies—exactly the AI application layer that’s struggling.
Major funds like Ares and Apollo have started *gating withdrawals*, blocking investors from pulling money out. Some are getting less than half of what they requested.
The problem? About $1.2 trillion in leveraged debt matures between 2027-2029. If rates stay higher, refinancing becomes brutal for borrowers barely covering interest payments today.
The Bottom Line
Three separate stories. One message: the easy-money assumptions that powered the last few years are getting stress-tested. How they hold up will separate the winners from the wishful thinkers.
