The Deleveraging of the AI Trade: What Goldman Sachs Misses About the Crypto Parallel

The silence between the digits holds the truth. On August 23rd, Goldman Sachs released a note that sent a quiet tremor through the market—not because of a crash, but because of a recalibration. The high-beta momentum portfolio fell 12% in a week. The AI hedge basket dropped 10% in five days. Leverage, that invisible architecture of modern finance, was being dismantled in real time. And yet, the headline read: "The AI trade is not over." This is the paradox of the current cycle. We built castles on the tidal data of sentiment, and now the tide is pulling back—not to retreat, but to reveal what was always beneath the surface. Goldman's analysis is not about AI. It is about the lifecycle of every technology narrative that has ever touched a ledger, from the dot-com boom to the DeFi summer. The question is not whether the trade is over. The question is whether we are willing to read the silence between the digits. Let me place this in the context of global liquidity. Since 2023, the AI trade has been a liquidity sponge, absorbing the excess reserves that central banks pumped into the system during the post-pandemic normalization. The M2 money supply, that tidal force of global markets, found its most concentrated expression in semiconductor equities and AI infrastructure. But liquidity is a ghost that haunts the ledger—it appears where you least expect it, and it vanishes when the narrative shifts. Goldman's note confirms that the ghost is moving. The capital that once flowed indiscriminately into AI hardware is now being redirected into storage, data centers, and—significantly—into traditional value sectors like European banks, gold miners, and copper producers. This is not a rotation. It is a structural recognition that the first phase of the AI trade—the phase where beta reigned supreme—has ended. The archive remembers what the algorithm forgets: every technology cycle follows the same arc. First, the infrastructure buildout, where capital is deployed without regard for revenue. Then, the reckoning, where the market demands proof of profit. Finally, the consolidation, where only the strongest survive. Goldman is telling us that AI has entered the second phase. The same arc applies to crypto, and this is where the parallel becomes uncomfortable. In crypto, we have been living through our own AI trade. The narrative shifted from "peer-to-peer electronic cash" to "digital gold" to "institutional adoption"—each phase attracting a new wave of leverage and speculation. The spot Bitcoin ETF approval in January 2024 was our equivalent of the AI infrastructure buildout. It brought Wall Street's liquidity into the asset class, but it also brought Wall Street's expectations. The ETF is not a tool for Satoshi's vision; it is a vehicle for the same momentum-driven, beta-chasing strategies that Goldman is now unwinding in AI. The question that keeps me awake at night is whether crypto's deleveraging is already underway, or whether it is merely delayed. Based on my experience auditing cross-border liquidity models for a Sydney-based bank in 2017, I learned that regulatory capital requirements are always one step behind the market. When I flagged the systemic risk of Bitcoin's volatility, management dismissed it as speculative novelty. That dismissal taught me something important: the market's memory is short, but its leverage is long. The same dynamic is playing out now. Goldman's note identifies storage and data centers as the most tactically attractive sectors because their profit recovery has not yet been reflected in stock prices. This is a classic signal of the second phase—the market is beginning to differentiate between companies that generate revenue and companies that merely tell stories. In crypto, the equivalent of storage and data centers is the infrastructure layer that most retail investors ignore. I am talking about the settlement layers, the oracle networks, the data availability solutions, and the energy infrastructure that powers proof-of-work networks. These are the sectors where profit recovery is real but underpriced. The archive remembers what the algorithm forgets: the value of a network is not in its token price, but in its ability to settle transactions without intermediaries. The market is beginning to understand this, but the understanding is uneven. Goldman's note also reveals a significant shift in the momentum factor. Software has replaced semiconductors as the largest weight in the three-month momentum long portfolio, while semiconductors have entered the short portfolio. This is a quant-level signal that the market's perception of value creation is moving from the "picks and shovels" of AI—the hardware—to the "gold miners"—the applications. In crypto, we are seeing the same shift. The momentum is moving from Layer-1 protocols and GPU-related tokens to application-layer projects that demonstrate actual user adoption. The real difference between OP Stack and ZK Stack is not technical—it is who can convince more projects to deploy chains first. The same logic applies to AI applications: the winner is not the one with the best model, but the one with the most users. But here is where I must introduce the contrarian angle. Goldman's analysis, for all its sophistication, misses a critical point: the AI trade and the crypto trade are not the same animal. AI is a productivity tool that generates revenue through enterprise adoption. Crypto is a monetary experiment that generates value through network effects and trust. The deleveraging of the AI trade will not necessarily trigger a deleveraging in crypto, because the underlying drivers are different. However, the risk is that the market treats them as the same asset class—a risk that has already materialized in the correlation between Bitcoin and the Nasdaq. We measured the shadow, mistaking it for the form. The shadow is the correlation between crypto and tech equities. The form is the fundamental difference between a decentralized monetary network and a centralized AI infrastructure. If the market continues to treat them as the same, then the AI deleveraging will drag crypto down with it. But if the market begins to differentiate, we may see a decoupling that creates the most significant alpha opportunity of this cycle. The transaction is cold; the trust is warm. This is the core insight that Goldman's analysis overlooks. AI is a cold transaction—it is about efficiency, computation, and optimization. Crypto is a warm trust—it is about consensus, decentralization, and human coordination. The deleveraging of the AI trade is a cold event. It will not change the fundamental warmth of crypto's value proposition. But it will test whether the market can distinguish between the two. Looking forward, the key signal to watch is not the price of Bitcoin or the earnings of Nvidia. It is the behavior of the momentum factor. If software continues to outperform semiconductors, and if crypto applications begin to outperform crypto infrastructure, we will know that the market is entering the second phase of the cycle—the phase where profit recovery matters more than narrative. The silence between the digits will tell us everything we need to know. Structure cannot contain the chaos of human hope. The AI trade is a structure built on hope—hope that the technology will transform every industry, hope that the revenue will eventually materialize, hope that the leverage will not unwind catastrophically. Crypto is a structure built on a different kind of hope—hope that a decentralized alternative to the current financial system is possible. Both structures are now being tested. The question is not which one will survive. The question is which one will emerge from the deleveraging with its core value proposition intact. I have spent the last six weeks in the Blue Mountains, disconnected from the noise, analyzing the monetary policy transmission mechanisms that connect the AI trade to the crypto trade. The conclusion is uncomfortable: the deleveraging is not over. The high-beta momentum portfolio has further to fall. The AI hedge basket has more pain to absorb. But the long-term trend remains intact. The key is to position for the second phase—the phase where profit recovery, not narrative, drives returns. In crypto, this means looking at the sectors that Goldman would identify as "tactically attractive" if it applied its framework to digital assets. Storage and data centers have their crypto equivalents: decentralized storage networks, data availability layers, and the physical infrastructure that powers the network. These are the sectors where profit recovery is real but underpriced. The market is beginning to understand this, but the understanding is uneven. The future is not a continuation of the past. It is a reconfiguration of the present. The AI trade is being reconfigured from a beta trade to an alpha trade. The crypto trade is being reconfigured from a retail trade to an institutional trade. Both reconfigurations are painful, but both are necessary. The silence between the digits holds the truth: the market is not ending. It is maturing. And maturity, as always, comes at a price.