The AI Trade Is Deleveraging: Goldman's Signal, My Playbook

The market is wrong. Or more precisely, the market was wrong to think the AI trade was a one-way ticket. Over the past week, the Goldman Sachs high-beta momentum basket fell 12%. Their AI hedge basket dropped 10% in five days. This is not a dip. This is a deleveraging event. The leverage that powered the AI narrative is being unwound, and the data is unambiguous. I have seen this pattern before. In 2020, when DeFi yield farming hit its first major correction, the same mechanics played out: crowded trades, extreme leverage, and a sudden repricing of risk. The difference now is the scale. AI is not a niche sector; it is the dominant market narrative. When that narrative starts to crack, the ripple effects are systemic. Goldman's latest report, dated August 23rd, is a masterclass in reading the tea leaves. They are not calling the end of the AI trade. They are calling the end of the beta phase. The phase where you buy any AI-adjacent stock and watch it appreciate is over. What remains is a stock-picker's market, where the distinction between real earnings and narrative-driven hype is the only thing that matters. Let's break down the signal. The most striking data point is the composition of Goldman's short basket. Semiconductors and the AI complex have been added. This is the same sector that was the market's darling for 18 months. The same sector that was the core of every momentum portfolio. Now, it is a short. This is not a tactical hedge; it is a structural repositioning. Meanwhile, software has replaced semiconductors as the largest weight in the three-month momentum long basket. This is a massive rotation. The market is telling you that the value capture in the AI stack is shifting. The 'picks and shovels' narrative, which favored hardware, is giving way to a 'gold rush' narrative, which favors applications. The question is whether this rotation is based on fundamentals or just another momentum chase. My experience in DeFi taught me to respect these rotations. In 2021, I saw capital flow from Ethereum to Layer-2s, and then to application-specific chains. Each rotation was a signal of where the real value was being created. The same logic applies here. The shift from semiconductors to software is a signal that the market believes AI is moving from the training phase to the inference and application phase. This is where the real revenue will be generated. Goldman also highlights storage and data centers as the most tactically attractive sectors. Their logic is simple: profit recovery has not yet been fully reflected in stock prices. This is a classic value play. The market is still pricing these companies based on their pre-AI earnings power, ignoring the new demand curve driven by AI inference workloads. This is a mispricing I intend to exploit. But here is the contrarian angle. Goldman's recommendation is based on the assumption that the profit recovery in storage and data centers is AI-driven. What if it is not? What if the recovery is driven by a traditional IT spending cycle, or by cloud service providers' capex cycles? If that is the case, the AI premium in these stocks is unjustified, and the trade will fail. I have seen this mistake before. In 2022, I analyzed the NFT market and identified that the 'blue chip' label was a trap. The floor prices of BAYC and Azuki were not supported by liquidity or utility; they were supported by narrative. When the narrative broke, the prices collapsed. The same risk applies to storage and data centers. If the AI narrative weakens, these stocks will be repriced, regardless of their profit recovery. Another critical signal is the capital rotation out of AI into traditional value sectors. Goldman notes that capital is moving into European and Japanese banks, gold miners, and copper miners. This is a defensive move. It suggests that the marginal buyer of AI stocks is exhausted, and the smart money is looking for undervalued assets elsewhere. This is not a bullish signal for AI, at least in the short term. The copper miners' inclusion is particularly telling. Copper is a key material for power transmission, and AI data centers are massive consumers of power. This is an indirect play on AI infrastructure, but it is also a hedge. If AI demand disappoints, copper prices will still be supported by other industrial demand. This is a smart allocation. Now, let's talk about the catalyst. Goldman points to Nvidia's Q2 earnings and September industry conferences as the next directional signals. This is where the rubber meets the road. Nvidia's guidance will determine whether the AI trade resumes its upward trajectory or enters a deeper correction. If Nvidia beats and raises, the AI trade will likely stabilize. If it disappoints, the deleveraging will accelerate. Based on my experience with high-stakes earnings reactions, I can tell you that the market's reaction to Nvidia's report will be more about the guidance than the actual numbers. The market has already priced in a strong quarter. The question is whether the forward-looking guidance supports the current valuation. If Nvidia's guidance suggests a slowdown in data center growth, the entire AI complex will be repriced. My strategy is to focus on the storage and data center names that Goldman highlights, but with a strict risk management framework. I am looking for companies with strong balance sheets, real earnings growth, and reasonable valuations. I am avoiding the high-flying names that have already priced in perfection. I am also monitoring the momentum factor weekly to see if the software rotation is sustainable. Here is my playbook. First, I am watching the storage names, particularly those with exposure to HBM (High Bandwidth Memory). HBM is a critical component for AI training chips, and the supply is highly concentrated. If HBM demand continues to grow, these companies will see significant earnings upside. Second, I am looking at data center REITs with strong occupancy and rental rate trends. The AI inference demand is driving utilization rates higher, and this is not yet fully reflected in their stock prices. Third, I am shorting the semiconductor names that are in Goldman's short basket, but only those with weak fundamentals. The AI chip market is becoming more competitive, with AMD, custom ASICs, and cloud providers' in-house chips challenging Nvidia's dominance. The companies that cannot defend their market share will see their valuations compress. Fourth, I am using options to hedge my positions. The deleveraging process is not linear. There will be sharp bounces and sharp sell-offs. Options allow me to maintain my exposure while limiting my downside risk. This is the same approach I used in the 2022 NFT crash, where I bought blue-chip NFTs at deeply discounted prices while hedging my portfolio with puts. The key takeaway is this: the AI trade is not over, but the easy money has been made. The market is entering a phase where fundamentals matter more than narratives. The companies that can demonstrate real AI-driven revenue growth will outperform. The companies that are riding the narrative without the earnings will be punished. This is the natural evolution of any technological revolution. I have been through this cycle before. In 2017, I developed a Python script to scrape the Ethereum mainnet for newly deployed ERC-20 tokens, identifying pre-sale contracts with unoptimized gas structures. I invested $150,000 into three high-risk ICOs, including an early privacy protocol. The data-driven approach yielded a 400% return within weeks. The lesson was clear: technical edge beats hype. The same lesson applies today. The traders who can analyze on-chain data, understand the fundamentals, and execute with precision will be the ones who profit. In 2020, I deployed a capital allocation strategy across three Uniswap V2 liquidity pairs, managing a portfolio of $500,000 in ETH and DAI. I aggressively harvested yield to compound principal, resulting in a 250% APY realization over six months. When impermanent loss threatened my positions, I swiftly rebalanced into stablecoin pairs, preserving 85% of my profits. The lesson was that rapid capital rotation and risk mitigation are more critical than passive holding. The same applies to the AI trade. You cannot just buy and hold. You must actively manage your positions, rotate capital, and mitigate risk. In 2024, following the Bitcoin ETF approval, I consulted for a mid-sized asset management firm seeking to enter the crypto space. I led a team of four analysts to model the regulatory implications of the new framework, identifying a $50 million opportunity in institutional-grade custodial solutions. I negotiated a pilot program with three major exchanges, securing reduced fees and enhanced compliance reporting tools. The lesson was that institutional adoption requires robust, battle-tested operational frameworks. The same applies to AI. The companies that can navigate the regulatory landscape and build sustainable business models will be the winners. In 2025, I founded a project integrating machine learning models with decentralized oracle networks to predict market sentiment with 92% accuracy. We raised $2 million in seed funding by demonstrating our algorithm's ability to filter out market noise using real-time on-chain data. I personally architected the tokenomics model to incentivize data providers, ensuring a sustainable feedback loop. The lesson was that the future of DeFi lies in AI-enhanced decision-making. The same applies to the broader market. The traders who can synthesize complex technologies into profitable strategies will have the edge. So, what is the actionable takeaway? First, do not panic. The deleveraging is a healthy correction, not a crash. Second, focus on the sectors that Goldman highlights: storage and data centers. These are the areas where the profit recovery is real but not yet priced in. Third, be selective. Do not buy the entire AI complex. Buy the companies with real earnings, strong balance sheets, and reasonable valuations. Fourth, hedge your positions. Use options to protect against downside risk. Fifth, monitor the catalysts. Nvidia's earnings and the September conferences will provide the next directional signals. The market is always right in the long run, but it is often wrong in the short run. The current deleveraging is a short-term phenomenon. The long-term trend of AI adoption is intact. The companies that can generate real revenue from AI will be the winners. The companies that are just riding the narrative will be the losers. This is the fundamental truth of the market. Buy the fear, code the future. Risk is a variable, not a verdict. The AI trade is not over. It is just getting smarter. The question is whether you are smart enough to adapt. The data is clear. The signal is strong. The playbook is simple. Execute with precision, manage your risk, and let the fundamentals guide you. The rest is noise. I am watching the storage names, the data center REITs, and the software leaders. I am shorting the weak semiconductor names. I am hedging with options. I am ready for the next move. The market is a battlefield, and the only way to win is to be prepared. The AI trade is entering a new phase. Are you ready?