When Machines Unwind: A Conversation with February's Chaos
When Machines Unwind: A Conversation with February's Chaos
The elevator pitch for what happened this week is easy enough to tell at a dinner party: commodities crashed, stocks rallied, crypto bled out. The Russell 2000 up 1%. Micro-cap chips running. Palantir beating on revenue. Manufacturing data robust. VIX at 17.44 and edging higher. The dollar had its best two-day move since April. Oil plunged on geopolitical optimism—Trump talking to Iran, suddenly the Middle East doesn't feel like a powderkeg anymore.
But the real story isn't in any single number. It's in what we don't say when we're reading the headlines.
The narrator in my head, the one who sits in a lot of calls with risk managers, kept asking: Where does a 10% gold selloff in 48 hours come from?
Not gradually. Not on bad data. Not on logical reassessment of where yields should be. But as a break.
Silver down 31% since Friday's close. An intraday plunge that wasn't seen since March 1980—44 years ago—when Volcker was strangling inflation with his bare hands and the Hunt brothers were trying to corner the market with borrowed money.
The parallels are sloppy, but the feeling is correct: something that people believed with absolute conviction stopped being believed. Instantly.
This is how markets actually move. Not via rational updating. Via collective unraveling.
The question I can't shake, though: What were these people actually long?
Precious metals didn't rally because the fundamental case for gold improved. Gold rallied because everyone on Earth had decided the Fed was about to cut rates aggressively. That became orthodoxy sometime around mid-January. Buy bonds. Buy gold. Buy anything that benefits from lower real rates. It was so obvious it barely needed defending. Rate cuts = asset prices up. This is the script we'd been running for 17 years.
Then Kevin Warsh gets nominated. Powell's independence gets questioned. Treasury yields move up 20 basis points. The narrative inverts in a weekend.
And people scramble out.
They scramble because they were never really holding gold on the merits. They were riding the momentum of a belief. When the belief broke, the position became indefensible. No amount of "but gold is a inflation hedge" helps when you bought at $5,550 expecting it to hit $6,000.
This is the part of markets they don't teach you in textbooks—the way structural positions can evaporate not because fundamentals changed, but because the temperature of hope changed.
Meanwhile, in equities, something else happened altogether.
The stock market did what it usually does when commodities crater: it bought the dip. The S&P 500 bounced 0.54%. The Dow climbed 1.05%. The Nasdaq took a baby step up 0.56% by Tuesday. This is the script in the other direction: fewer rate cuts means better earnings growth because the cost of capital stabilizes and recession risks recede. Companies can stop apologizing for capex and start talking about returns on capital.
Micron (MU) surged 5.52%. AMD soared 3.95%. Palantir beat on revenue growth. Raymond James upgraded Alphabet (GOOG) to strong buy with a $400 price target because—get this—they think Google's AI stack is "shifting to high gear." Which is a wonderful way of saying: We think the company spent money wisely and it's about to pay off.
But here's where my narrator gets nervous.
The narrative that drove equity outperformance the past 18 months was "AI is worth whatever we say it's worth because growth transcends profitability during paradigm shifts." That gets harder to defend when your central thesis depends on rates staying low to justify astronomical capex spending. It gets much harder when the Fed chair nominee has spent a career arguing that low rates breed complacency and asset price bubbles.
So Nvidia, naturally, got clobbered on news that its $100 billion investment in OpenAI was stalled. The market felt the contradiction even before anyone articulated it. Companies that gorged on cheap leverage to fund speculative AI infrastructure are going to look very different when the cost of capital actually means something.
Bitcoin's journey this week was its own kind of story. A fable, really.
From $83,000 in January to $76,827 by Monday to $75,000 by Wednesday to brief touches of $74,000 over the weekend. The fourth consecutive monthly decline. The longest losing streak in seven years. Bitcoin is down 14% year-to-date. Every major and minor altcoin is bleeding roughly the same amount, if not more.
This is the moment when crypto insiders get angry at the market for not understanding that Bitcoin has "emerged from its discount window" and is now "institutional-grade." And maybe that's true. Or maybe Bitcoin is just learning, again, that it moves in lockstep with the macroeconomic regime. When rates were supposed to be permanently low, Bitcoin was a hedge. When rates might actually stay higher, Bitcoin is leverage.
Leverage blows up.
Hyperliquid (HYPE) was up 10%—a lone bright spot in a field of carnage—because it's built on tokenized gold and silver trading and got a bid from people rotating out of physical precious metals into instruments they can margin. That's not a fundamental breakthrough. That's parasitic activity on top of a larger unwinding.
The crypto cognoscenti kept saying: "This is capitulation. Capitulation is bullish." And maybe they're right. But I kept thinking: Capitulation on what basis? Capitulation because you finally accepted that rate cuts aren't coming? Or capitulation because you got margin-called and had to sell at any price?
One of those becomes a bottom. The other just becomes a crater with a floor somewhere lower.
The real question nobody wants to ask yet is whether the equity market's confidence is justified or just another narrative waiting to break.
Yes, manufacturing data is solid. January ISM at 52.6, the best since 2022. New orders rising. Production expanding. But look underneath: unemployment is stable but job growth is sluggish—only 50k adds in December, down sharply from trend, with downward revisions of 76k from the prior two months.
Wages are growing, but mostly for supervisory and high-skilled roles. Production and non-supervisory workers—roughly 80% of private payrolls—saw wage growth slow to 3.6% year-over-year. The labor market isn't weak. But it's not screaming strength either.
So here's the version of reality where the stock market is wrong: tariff pass-through hits harder than expected. Companies can't pass price increases to consumers. Margins compress. The Fed stays on hold not out of Hawks confidence, but out of genuine uncertainty. The Fed under Warsh holds rates steady while inflation stays sticky. Oil rebounds as geopolitical premium returns. Manufacturing stumbles into a soft data surprise.
And here's the version where the market is right: the Fed's caution creates a data-dependent path toward eventually lower rates, but only after we get enough evidence that inflation is tamed. Companies with fortress balance sheets outperform. AI capex pays off because it had to. Mean reversion favors equities against commodities. The malaise of 2025 gives way to a rational 2026.
One of these narratives will anchor the next six months. We won't know which until March.
This is where I'm supposed to tell you what to do.
I won't. Because the honest thing is this: if you'd held through the Warsh shock and the gold implosion and the crypto carnage, you'd have been right. The S&P 500 closed near all-time highs. The Dow printed 49,407. The Russell 2000, that overlooked proxy for Main Street growth, climbed 1%.
But that's backward-looking. The forward-looking question is whether the market repriced on new information or just cycled through a set of options until it landed on the one that felt least scary.
My narrator thinks the repricing was real. The Fed's policy path does matter more than people think. Lower leverage is healthy. The correction in precious metals probably needed to happen.
But my narrator also remembers 2018, 2020, and 2022—years when the markets seemed to settle and then didn't. Years when conviction proved to be a liability. Years when the thing everyone agreed on turned out to be the first domino to fall.
So here's what I actually believe: Hold your core positions. Stay alert to data that contradicts the current narrative. And keep asking yourself whether you're long because the fundamental case is sound, or because you're riding someone else's conviction.
Because when conviction breaks, everything else breaks with it.
February's volatility wasn't a glitch. It was a reset. Maybe it's a healthy one. Or maybe it's the first act of a longer unraveling.
We'll know by April. Until then, the machines are trying to figure out what to do with all this repositioned capital, and everyone else is staring at their portfolios asking the same question I've been circling all week:
Did the market reprice on fundamentals, or just find a new story to believe?
The answer matters. And it's not obvious yet.