The AI Trade's Second Act: Deleveraging, Divergence, and the Storage Signal

The high-beta momentum portfolio lost 12% in a single week. The AI hedge basket dropped 10% in five days. These are not crash numbers from a bubble bursting—they are the signature of a market unwinding leverage, and Goldman Sachs is telling us the trade is not over, only changed. As a digital asset fund manager who has watched liquidity cycles bend and break across crypto and equities, I recognize this pattern. It is the moment when the tide of beta recedes, and only the rocks of real earnings remain visible. Goldman's framing is precise: the AI trade has entered a deleveraging phase, but the structural opportunity has shifted from broad sector exposure to individual stock selection. The firm explicitly identifies storage and data centers as the most tactically attractive sectors, arguing their profit recovery is not yet reflected in share prices. This is a signal that the AI value chain is migrating from the training layer to the inference and data infrastructure layer. In my 2024 work integrating BlackRock's IBIT flow data into our Nairobi fund's liquidity models, I observed a similar transmission lag—institutional capital moves in waves, and the second wave often targets the picks-and-shovels of deployment, not the narrative of discovery. The most telling detail is the momentum factor rotation. Software has replaced semiconductors as the largest weight in the three-month momentum long portfolio, while semiconductors have entered the short basket. This is not a casual rebalancing; it is a quant-level verdict on where value is being captured. The market is signaling that the era of paying a premium for GPU scarcity is yielding to a focus on application-layer revenue and data infrastructure profitability. The ledger remembers what the algorithm forgets—and the algorithm is now remembering that storage and data centers have pricing power, stable supply dynamics, and a demand curve steepening with every new AI inference request. Here is the contrarian angle most analysts will miss. The capital rotating out of AI into European and Japanese banks, gold miners, and copper stocks is not a retreat from technology. It is a hedge against the very real possibility that AI infrastructure buildout faces a power and materials bottleneck. Copper is the transmission metal for data center electricity; gold is the hedge against the monetary expansion that will fund the next wave of AI capex. Goldman is not saying AI is over. They are saying the trade is maturing, and maturity demands diversification. For those of us who lived through the Terra collapse and the 2022 deleveraging, the playbook is familiar. The first phase of any technological revolution is funded by narrative and liquidity. The second phase is funded by earnings and operational efficiency. Storage and data centers are the quiet beneficiaries of this transition. Their profit recovery is real, but the market's attention is still fixated on the next Nvidia earnings print. That divergence is the opportunity. Trust is borrowed; trust is never owned. The market's trust in the AI trade is being re-earned through fundamentals, not narratives. Safety is the only yield that compounds over time, and in this market, the safest position is in the sectors where earnings have already arrived but the stock price has not yet caught up. The question is not whether AI will continue to reshape the global economy. The question is whether you are positioned for the second act, where the story is written in storage capacity, data center utilization, and the quiet compounding of infrastructure profits. We build walls not to keep out, but to keep safe—and in this market, the walls are built of earnings visibility and balance sheet discipline.