Hook: The Index That Shouldn't Exist Is Outperforming.
Since June 2025, a hypothetical index—one that explicitly excludes the AI complex—has outperformed the standard S&P 500. Goldman Sachs built it. The data is unambiguous. This is not a blip. It is a structural signal that the market's primary growth engine has stalled relative to the rest of the economy.
I have spent the last 24 years reconciling on-chain data with market reality. When I see a divergence like this, I do not ask what it means for AI. I ask what it means for the other 450 companies in the index. The answer is more complex than the mainstream narrative suggests. This isn't a simple 'risk-on' rotation. It's a repricing of duration risk, a quiet admission that the AI trade is crowded, and a potential precursor to a broader liquidity event.
Context: Deconstructing the Goldman Non-AI Index.
The core fact is simple: Goldman Sachs constructed an S&P 500 variant that strips out AI-related equities. Since June 2025, this 'non-AI' index has outperformed the cap-weighted standard index. The timing is critical. June 2025 was the apex of the AI mania—Nvidia's market cap had just blown past $5 trillion, and a handful of AI names were printing all-time highs with mechanical regularity. The fact that the ex-AI index has outperformed since that exact peak suggests we are not looking at a random fluctuation. We are looking at the market pricing in a peak in marginal AI capital expenditure.
My background is in standardizing chaotic data—whether it was tracking 1,200 ICOs in 2017 or auditing flash loan activity on Aave in 2020. The first question I always ask when analyzing a new index is: what is the selection bias? Goldman hasn't published the full methodology, but the implication is clear. The index likely underweights or eliminates the 'Magnificent Seven' and semiconductor names. This creates a bias toward value, industrials, financials, and consumer discretionary.
This is where the data gets interesting. The outperformance is not just a 'value vs. growth' story. It is a 'duration vs. cash flow' story. AI stocks are priced on optionality and future earnings that may be years away. Non-AI stocks are priced on current earnings. When the non-AI index outperforms, it signals that the market's marginal buyer prefers immediate, verifiable cash flows over speculative future potential. This is the behavior of a market that is starting to fear the discount rate.
Core: The On-Chain and Macro Evidence Chain.
Let's quantify the manipulation of the narrative. The prevailing market story is that AI is the only game in town. But the price action suggests otherwise. We need to look at this through the lens of capital flows.
First, consider the 'Yield Farming' analogy. In DeFi, when a protocol offers 200% APY, it attracts liquidity—until the incentives run out. The AI trade has been the ultimate yield farm. The 'yield' was massive capital gains, but the underlying 'TVL' (Total Value Locked) is the actual earnings power of these companies. If we strip out the narrative, the question becomes: are AI companies generating cash flows that justify their valuations? Based on the price action since June, the market's answer is 'no' or 'not yet.'
Second, look at the breadth data. The outperformance of the ex-AI index implies that market breadth is improving. This is a classic late-cycle signal. In the early stages of a bull market, leadership is narrow. In the late stages, it broadens as money rotates into laggards. The fact that the non-AI index is winning is a sign that the 'easy money' in the AI trade has been made. Follow the gas, not the hype. The 'gas' here is the rotation of capital from high-beta tech into lower-beta industrials and financials.
Third, we must analyze the interest rate correlation. AI stocks are long-duration assets. Their valuations are highly sensitive to the discount rate. If the market anticipates 'higher for longer' interest rates, these long-duration assets face downward pressure. Conversely, value stocks in the non-AI index have shorter duration profiles. They are less sensitive to rate fluctuations. The outperformance of the non-AI index is consistent with a market that is repricing the path of monetary policy. It suggests that traders are no longer betting on aggressive rate cuts that would fuel speculative growth.
This is the core insight: the market is not saying 'AI is bad.' It is saying 'AI is expensive.' The distinction is critical. AI remains a transformative technology, but the price-to-earnings ratios across the sector have detached from reality. The non-AI index's outperformance is a correction mechanism, forcing capital back into sectors where the earnings yield is more attractive.
Contrarian: The Correlation Trap—Why This Is Not a 'Broadening' Rally.
Here is where my forensic skepticism kicks in. The mainstream interpretation of this data will be positive: 'The rally is broadening! The economy is strong! It's not just tech!' I reject that conclusion.
This is not a broadening rally. This is a defensive rotation disguised as diversification. When money flows from high-beta tech into low-beta staples and utilities, it is not a sign of confidence; it is a sign of fear. The non-AI index includes defensive sectors that outperform during uncertainty. If this were a true 'growth diffusion' signal, we would see cyclical sectors like materials and energy leading. Instead, we need to look at the quality of the earnings in the non-AI index. Are they beating estimates because of organic growth, or because of cost-cutting and buybacks? DeFi efficiency is math, not marketing. The same applies to the S&P 500. The math suggests that ex-AI earnings growth is anemic, and the relative outperformance is due to a contraction in AI multiples, not an expansion in ex-AI multiples.
Furthermore, consider the 'CryptoPunks' wash trading pattern. In 2021, I audited NFT floor prices and found that 15% of 'reported' volume was fake. The market was looking at a distorted picture. The same distortion exists in the equity market. The 'AI narrative' has been pumped by a coordinated chorus of sell-side analysts and media. The Goldman 'non-AI' index is a contrarian tool. It is a way to hedge against the concentration risk that has built up in the 'AI' trade. Goldman is not publishing this index because they believe in diversification. They are publishing it because they see the same risk I see: the AI trade is a liquidity bubble, and when it pops, it will take the entire index down with it.
The correlation trap is assuming that 'non-AI' equals 'safe.' It does not. If the AI bubble bursts, the liquidity crunch will hit all risk assets. The non-AI index will outperform on a relative basis, but it will still lose money on an absolute basis. Investors who rotate into the non-AI index thinking they are 'safe' are making a classic error. They are buying a lifeboat on a ship that is still taking on water.
Takeaway: The Signal to Monitor for the Next 90 Days.
The data since June tells me that the marginal buyer is exhausted with AI. They are looking for yield, safety, and cash flow. But the sustainability of this rotation is fragile.
I am tracking three specific data points. First, the Copper-to-Gold ratio. If the non-AI outperformance is driven by genuine cyclical recovery, copper prices should be rising relative to gold. If gold is outperforming copper, this is a risk-off signal, and the non-AI outperformance is purely defensive. Second, I am watching the 10-year Treasury yield. If yields rise alongside the non-AI outperformance, it confirms a 'growth diffusion' narrative. If yields fall, it confirms a 'flight to safety' narrative. Third, I am monitoring the next earnings season for the 'non-AI' names. They need to show actual revenue growth, not just margin expansion.
The most dangerous scenario is a 'taper tantrum' in the AI sector. If a major AI player issues weak guidance or announces a cut in capital expenditure, the rotation could reverse violently. Capital would flee back into the safety of the mega-caps, and the non-AI index would underperform.
My verdict is simple. This is not a time for celebration. It is a time for hedging. The non-AI index outperformance is a warning sign that the market is losing its primary catalyst. Data doesn't lie, but narratives do. The narrative of 'broadening participation' is a lie. The reality is a market seeking cover from its own excesses. Standardize your risk metrics, reduce leverage, and do not mistake a defensive rotation for a fundamental shift in economic growth. The next few quarters will determine whether this is a pause or a pivot. The data suggests we are closer to the latter than the former.
