The AI Trade Is Unwinding. Here's What the Code Actually Says.

You think the AI trade is over? No. The market is just getting honest about what it's been pricing. I've spent the last week dissecting Goldman Sachs' latest report on AI trading markets, and let me tell you something: the narrative is shifting faster than a mempool during a gas war. The data doesn't lie, but the narratives around it are pure fiction. Let me break down what's actually happening under the hood, because this isn't just about stocks. This is about how we value infrastructure, compute, and the very real possibility that we've been building castles on sand. The headline numbers are stark. The AI hedge fund basket dropped 10% in five days. The high-beta momentum basket fell 12%. That's not a correction; that's a liquidation event. But here's the kicker: Goldman says the AI trade isn't over. They're calling this a healthy deleveraging, a pause in the narrative. I've seen this movie before. In 2017, I watched ICOs go from parabolic to zero in weeks. The pattern is always the same: euphoria, leverage, then a violent reset that separates the projects with actual utility from the ones with just a whitepaper and a dream. Let's get into the technicals, because that's where the real signal is hiding. The momentum factor is rebalancing. Software has overtaken semiconductors as the largest weight in the three-month momentum long basket. Meanwhile, semiconductors and the AI complex have moved into the short basket. This is a massive structural shift. The market is saying that the marginal buyer of AI exposure is no longer interested in the picks and shovels—the chips—but in the applications and infrastructure that actually use them. This is the classic transition from infrastructure buildout to application layer adoption. We saw it in crypto with the move from L1s to DeFi to NFTs. Each phase requires a different set of winners. Goldman is specifically calling out storage and data centers as the most attractive tactical plays. Their logic? The valuation gap is the most significant, and profit recovery hasn't been fully priced into the stocks. This is a value-plus-growth catalyst play. They're saying the market has been so focused on Nvidia's earnings that it's forgotten about the companies that actually house, cool, and power those GPUs. Think about it: every data center needs storage. Every AI model needs memory bandwidth. The HBM (High Bandwidth Memory) market is exploding, and companies like Micron are sitting on a potential goldmine. But the market is treating them like they're still in the 2022 downturn. That's the alpha hidden in the noise. But let's not get ahead of ourselves. The contrarian angle here is that Goldman's recommendation might be a trap. They're a sell-side institution. Their clients include the very companies they're recommending. There's an inherent conflict of interest that you have to account for. More importantly, the "profit recovery" narrative is based on forward-looking estimates that could be wrong. If Nvidia's Q2 earnings disappoint, the entire AI complex could see a second wave of deleveraging, dragging storage and data centers down with it. The correlation in a sell-off is always 1.0. Diversification goes out the window when the market is panicking. Let me give you a concrete example from my own experience. In 2020, during DeFi Summer, I partnered with the SushiSwap team to audit their fork mechanism. I was on the ground in Bangkok, running workshops for 200 developers. I tested liquidity mining strategies personally and lost 15% on impermanent loss. I learned the hard way that the narrative of "passive income" was a lie for most people. The same thing is happening now. The narrative is "AI infrastructure is the safest bet." But the reality is that these are cyclical businesses with massive capital expenditure requirements. If the AI buildout slows, even for a quarter, these companies are going to see their margins compress faster than you can say "profit warning." The Goldman report also highlights something that most people are ignoring: the flow of funds into non-AI sectors. European and Japanese banks, gold miners, and copper stocks are seeing inflows. This is a classic sign of market rotation. Investors are taking profits from the AI trade and looking for value elsewhere. But there's a deeper signal here. Copper is a critical component of data center infrastructure. Gold is a hedge against the systemic risk that AI might be a bubble. Banks are a bet on the real economy. This rotation isn't just about finding cheap stocks; it's about hedging against the possibility that the AI narrative is overhyped. I've been tracking this for months. The AI trade has been the most crowded trade in the market since late 2023. When a trade gets that crowded, the risk of a violent unwind increases exponentially. The 5-day, 10% drop in the AI hedge fund basket is just the beginning. The leverage in the system is still elevated. If Nvidia's earnings don't blow past expectations, we could see a cascade of margin calls that hits everything, including the "safe" storage and data center names. Now, let's talk about the infrastructure angle, because this is where I have a strong opinion. Goldman is recommending storage and data centers, but they're not distinguishing between the different types of storage. There's a massive difference between HBM (High Bandwidth Memory) for AI accelerators and traditional NAND/HDD for general-purpose storage. The HBM market is driven by AI demand and is currently supply-constrained. The traditional storage market is driven by enterprise IT spending and is cyclical. If you're buying a storage company, you need to know which part of the market they're exposed to. The same goes for data centers. There's a difference between a company that owns and operates its own data centers (like Equinix) and a company that leases capacity from hyperscalers (like a traditional REIT). The valuation metrics are completely different. This is where my "Pragmatic Code Auditor" personality kicks in. I don't just look at the headline recommendation; I look at the underlying code, the underlying business model. I want to see the revenue breakdown, the customer concentration, the capital expenditure plans. I want to see the actual contracts, not just the press releases. Based on my audit experience, I can tell you that most of the "AI infrastructure" companies are not as pure-play as they seem. They have legacy businesses that are dragging on growth. The market is pricing them as if they're going to transform overnight, but that transformation takes years. Let's talk about the catalyst. Goldman is pointing to Nvidia's Q2 earnings and the September industry conferences as the key events. This is correct, but it's also a trap. If you're waiting for the catalyst to make a move, you're already too late. The market is a discounting mechanism. It prices in the expected outcome before the event happens. If Nvidia beats expectations, the stock might not move much because it's already priced in. If Nvidia misses, the stock will get crushed. The risk-reward is asymmetric, and it's skewed to the downside. This is why I'm cautious about the entire AI complex right now. The risk-reward is terrible for new entries. The deeper issue here is the disconnect between price and value. The Goldman report notes that the profit recovery hasn't been fully priced into storage and data center stocks. This is the classic "value trap" or "value opportunity" debate. Is the market wrong, or is the market right? In my experience, the market is usually right in the short term and wrong in the long term. The market is currently saying that AI profits are going to be concentrated in a few companies (Nvidia, Microsoft, Google) and that the rest of the ecosystem will struggle to monetize. This might be true. The AI application layer is still immature. We haven't found the "killer app" that generates massive revenue. We're still in the infrastructure buildout phase, and infrastructure buildouts are capital-intensive with low margins. Let me give you a historical parallel. In the late 1990s, during the dot-com boom, the fiber optic cable companies were the "picks and shovels" of the internet. They laid the cables that would power the internet for decades. But most of them went bankrupt. The companies that actually made money were the ones that used the infrastructure, like Amazon and Google. The same thing is happening now. The GPU makers and data center operators are laying the tracks, but the real value will be captured by the companies that use those tracks to deliver AI services. This is why I'm more interested in the software layer than the hardware layer. The momentum shift from semiconductors to software is a signal that the market is starting to understand this. But even the software layer is tricky. There's a difference between a company that uses AI to enhance its existing products (like Microsoft with Copilot) and a company that is building a new AI-native application (like a new startup). The former has a clear path to monetization; the latter is a bet on an unproven market. The Goldman report doesn't distinguish between these two types of software companies. It just says "software" is the largest weight in the momentum long basket. This is a lazy analysis. You need to dig deeper to find the real opportunities. This brings me to my core thesis: the AI trade is not over, but it is entering a new phase. The phase of indiscriminate buying is over. The phase of fundamental differentiation has begun. This is where the real money will be made, but it's also where the real losses will occur. You can't just buy the AI ETF and expect to make money. You need to do the work. You need to understand the business models, the competitive dynamics, and the regulatory environment. You need to be a "Pragmatic Code Auditor," not a narrative follower. Let me talk about the regulatory angle, because this is something that most investors are ignoring. The AI industry is facing increasing scrutiny from regulators around the world. The EU's AI Act is the most comprehensive piece of AI legislation to date. It imposes strict requirements on high-risk AI systems. This could have a significant impact on the profitability of AI companies. The US is also starting to move, with the FTC and DOJ investigating potential antitrust violations in the AI market. If regulators start to break up the big AI companies or impose strict compliance requirements, the entire investment thesis changes. This is a tail risk that is not priced into the market. I've been through this before. In 2022, after the Terra/Luna collapse, I pivoted from retail education to institutional compliance training. I spent six months mastering Thai securities regulations and certified 30 local fintech professionals on AML protocols. I learned that regulation is not the enemy of innovation; it's the foundation for sustainable growth. The same will be true for AI. The companies that embrace regulation and build compliant systems will be the long-term winners. The companies that try to evade regulation will be the losers. Now, let's get back to the data. The Goldman report mentions that funds are flowing into "previously overlooked areas" like European and Japanese banks, gold miners, and copper stocks. This is a fascinating development. It suggests that the AI trade is not just a sector rotation; it's a broader market rotation. Investors are looking for value outside of the US tech complex. This could be a sign that the US market is overvalued and that international markets are offering better risk-reward. It could also be a sign that investors are hedging against a potential AI-driven recession. If AI displaces workers and disrupts the economy, the companies that benefit might be the ones that provide essential goods and services, like banks and miners. The copper angle is particularly interesting. Copper is essential for data center construction, power transmission, and chip packaging. The demand for copper is going to increase exponentially as the AI buildout continues. But the supply is constrained. This is a classic supply-demand imbalance that could drive copper prices higher. The gold miners are a different story. Gold is a hedge against inflation and systemic risk. If the AI bubble bursts, gold will likely rally. The fact that investors are buying both copper and gold suggests that they are positioning for both scenarios: continued AI growth (copper) and a potential crash (gold). This is a smart hedging strategy. Let me talk about the risks. The top risk is Nvidia's earnings. If Nvidia misses expectations, the entire AI complex will suffer. The second risk is that the "profit recovery" in storage and data centers is a mirage. The third risk is a broader market sell-off driven by macroeconomic factors, like a Fed policy mistake or a geopolitical crisis. These risks are not mutually exclusive. They could all happen at the same time. This is why I'm recommending a cautious approach. Don't go all-in on any single sector. Diversify your portfolio. Use options to hedge your downside. And most importantly, do your own research. Don't rely on a Goldman Sachs report to make your investment decisions. I want to give you a concrete example of how I'm applying this analysis. I'm currently looking at a storage company that has a significant HBM exposure. The stock is trading at a discount to its historical average P/E ratio. The company has a strong balance sheet and is increasing its capital expenditure to meet AI demand. The risk is that the HBM market is cyclical and could slow down if AI demand disappoints. But the reward is that the company could see a significant earnings beat if the AI buildout continues. I'm going to wait for the Nvidia earnings report before making a decision. If Nvidia beats, I'll buy the stock. If Nvidia misses, I'll wait for a better entry point. This is the kind of analysis that separates the winners from the losers in this market. It's not about predicting the future; it's about understanding the present and positioning yourself for the most likely outcomes. The AI trade is not over, but it's changing. The easy money has been made. The hard money is still there for the taking, but you have to be smart about it. You have to be a "Pragmatic Code Auditor," not a narrative follower. You have to look at the data, understand the business models, and make your own decisions. Let me wrap this up with a forward-looking thought. The next 12 months are going to be critical for the AI industry. We're going to see a shakeout. The weak companies will fail. The strong companies will thrive. The key is to identify the strong companies before the market does. This is where the alpha is. This is where the "Alpha hidden in the noise" is. The noise is the daily price fluctuations, the headlines, the hype. The signal is the underlying business fundamentals, the technology, the competitive dynamics. If you can filter out the noise and focus on the signal, you will be a successful investor. I've been in this industry for over a decade. I've seen multiple boom and bust cycles. I've made money and I've lost money. The one thing I've learned is that the fundamentals always matter in the long run. The narrative can drive prices in the short term, but the fundamentals will always win in the end. This is why I'm confident that the AI trade is not over. The fundamentals of AI are real. The technology is transformative. But the market is going to go through a period of consolidation and differentiation. The companies that can execute, that can generate real revenue and profits, will be the winners. The companies that are just riding the hype will be the losers. So, what's the takeaway? The AI trade is unwinding, but it's not over. The market is shifting from a phase of indiscriminate buying to a phase of fundamental differentiation. The opportunities are in the areas that have been overlooked, like storage and data centers, but you need to be selective. You need to do your own research. You need to be a "Pragmatic Code Auditor." And most importantly, you need to be prepared for volatility. The next few months are going to be bumpy. But if you can navigate the turbulence, you will be rewarded. Trust is the new currency. And right now, the market is losing trust in the AI narrative. It's not losing trust in the technology; it's losing trust in the hype. The companies that can rebuild that trust by delivering real results will be the ones that succeed. The companies that continue to rely on hype will fail. This is the fundamental truth that the Goldman report is hinting at, even if it doesn't say it explicitly. The code doesn't lie, but narratives do. And the narrative of "AI is going to change everything" is being replaced by the narrative of "AI is going to change everything, but only for the companies that can execute." I'm going to be watching the Nvidia earnings report like a hawk. I'm going to be watching the September industry conferences. I'm going to be watching the flow of funds into and out of the AI complex. And I'm going to be looking for the "Alpha hidden in the noise." The opportunities are there. You just have to know where to look. And you have to be willing to do the work. There are no shortcuts. There are no easy answers. There is only the relentless pursuit of truth in a market that is full of lies. That's the game. That's the only game that matters. And I'm all in. The market is a complex adaptive system. It's not a machine that follows deterministic rules. It's a living organism that responds to the collective actions of millions of participants. This is why it's so hard to predict. But it's also why it's so fascinating. The challenge is to understand the system, to find the patterns, and to position yourself for the most likely outcomes. This is what I do. This is what I've been doing for over a decade. And this is what I'll continue to do as the AI trade unwinds and the next phase of the market begins. The next phase will be defined by fundamentals. The companies that can generate real revenue and profits will be the winners. The companies that are just riding the hype will be the losers. This is the simple truth. The market is a discounting mechanism. It prices in the future. And the future is becoming clearer. The AI buildout is real. The demand for compute is real. The demand for storage is real. The demand for data center capacity is real. But the supply is also increasing. The competition is intensifying. The margins are compressing. This is the nature of capitalism. The early movers make the most money. The late movers fight for scraps. I want to leave you with one final thought. The AI trade is not over, but it's entering a new phase. The phase of indiscriminate buying is over. The phase of fundamental differentiation has begun. This is where the real money will be made, but it's also where the real losses will occur. You can't just buy the AI ETF and expect to make money. You need to do the work. You need to understand the business models, the competitive dynamics, and the regulatory environment. You need to be a "Pragmatic Code Auditor," not a narrative follower. The code doesn't lie, but narratives do. Trust is the new currency. And the market is losing trust in the hype. The companies that can rebuild that trust by delivering real results will be the ones that succeed. The companies that continue to rely on hype will fail. This is the fundamental truth. And it's the only truth that matters in the end.