Goldman's AI Trade Is Bleeding. The Smart Money Just Found the Exit.

The smell of fear is back in the AI trade. Over the past five days, Goldman's AI hedge basket has shed 10%. The high-beta momentum basket? Down 12%. This isn't a dip. It's a deleveraging event. And the narrative that 'AI is over' is already starting to circulate on every Discord server I'm in. But here's the thing I keep telling people: Algorithms smell fear, but they respect speed. The market isn't abandoning AI. It's just repricing it. And if you're still holding the same bags you bought in January, you're the exit liquidity for someone who read the momentum data faster than you did. Let's cut through the noise. Goldman Sachs just dropped a note that should be required reading for anyone in this market. Their core thesis is simple: the era of buying the entire AI sector and watching it go up is over. The 'rising tide lifts all boats' phase has officially ended. What we're seeing now is a violent rotation, not a capitulation. The data from their prime brokerage desk shows momentum factors are rebalancing in real-time. Software has overtaken semiconductors as the largest weight in the three-month momentum long basket. Meanwhile, semiconductors and the broader AI complex have been shoved into the short basket. That's not a typo. The market is now betting against the very chips that powered this entire cycle. This is the context most retail traders are missing. We've been conditioned to think of AI as a monolith. Nvidia goes up, everything goes up. But the market is now differentiating. Goldman is explicitly telling clients to look at storage and data centers. Their reasoning? The profit recovery in these sectors hasn't been fully priced into the stock prices yet. The valuation gap is the most significant they've seen. This is classic value-plus-catalyst logic. They're not saying AI is dead. They're saying the easy money has been made on the picks and shovels narrative, and now you have to find the companies where the earnings haven't caught up to the hype. I've seen this movie before. It's the same pattern we saw in the DeFi summer of 2020. Everyone was chasing the highest APY, the flashiest protocol, the biggest TVL. Then the music stopped, and the projects with real usage survived while the ones with just a pretty dashboard and a token dump went to zero. The AI trade is going through the same maturation process. The 'concept' phase is over. The 'delivery' phase has begun. And the market is brutally efficient at punishing those who don't adapt. Here's the contrarian angle that Goldman isn't shouting from the rooftops. The money flowing out of AI isn't just going into storage. It's rotating into European and Japanese banks, gold miners, and copper stocks. This is the part that should make you think. Copper. Why copper? Because AI data centers consume massive amounts of power, and that power needs physical infrastructure. The market is starting to price in the physical footprint of the AI buildout. This isn't just a financial rotation; it's a recognition that the AI supply chain extends far beyond TSMC and Nvidia. The 'AI+traditional industry' crossover trade is being born right now. But let's be clear about the risks. The deleveraging we're seeing might not be over. The AI hedge basket is down 10% in five days, but that's after a massive run-up. Leverage levels are coming down from extreme highs, but they might still be elevated. If Nvidia's Q2 earnings, which are due around August 28th, disappoint or offer weak guidance, we could see a second wave of selling. And that selling would drag down the storage and data center names that Goldman is recommending. The correlation in this market is still dangerously high when fear hits. My take based on years of watching these cycles: the storage and data center thesis is sound, but the timing is everything. Goldman is right that the profit recovery isn't priced in. But the market is a voting machine in the short term and a weighing machine in the long term. Right now, the votes are being cast by momentum algorithms that are scared. You need to be patient. Don't try to catch the falling knife. Wait for the Nvidia earnings to pass, let the market digest the news, and then look for entries into the names that have actual earnings visibility. Micron, Dell, Super Micro—these are the types of companies that will benefit from the infrastructure buildout, but they will also get caught up in the sector-wide selloff if Nvidia stumbles. We don't know if this is the bottom or just a pause before another leg down. But we do know that the market is telling us something. It's telling us that the 'buy everything AI' trade is dead. It's telling us that differentiation is the new alpha. And it's telling us that the next big opportunity might not be in the obvious places. Yield is a drug; exit liquidity is the cure. The smart money is taking profits on the crowded trades and positioning for the next phase. The question is, are you smart enough to follow them, or are you going to be the one holding the bag when the music stops? Chaos is just data waiting for a narrative. The narrative is shifting. Pay attention.