The market is not rational; it is resistant. When Chinese hedge funds begin to rotate out of Nvidia and the hyperscalers, calling AI a 'super bubble', the signal is not a whisper—it is a fracture. I have seen this pattern before. In 2017, during the ICO boom, I audited over 50 whitepapers for a Stockholm-based fund. The infrastructure was real, but the pricing had already discounted a decade of adoption. The same mechanics are at play here, only this time the ledger is global liquidity, and the code is the capital expenditure of the largest cloud providers.
Context: The Macro Map of Liquidity and AI
To understand the move, we must first map the global liquidity terrain. The Federal Reserve's rate hikes have tightened the supply of cheap capital, the lifeblood of high-growth narratives. Meanwhile, the capital expenditure of the four hyperscalers—Microsoft, Google, Amazon, and Meta—has crossed an annualized $200 billion, a figure that dwarfs the entire crypto market cap. The AI infrastructure narrative is built on the assumption that demand for compute will grow exponentially, but the cost of that compute is now being questioned. The Chinese hedge funds are not just selling a stock; they are selling a thesis that the marginal return on capital for AI infrastructure is collapsing.
This is not a new phenomenon. In the 2000s, telecoms laid fiber across the ocean, and the bubble burst before the bandwidth was fully utilized. In 2021, SaaS valuations cratered after a similar overinvestment cycle. The 'super bubble' label is not hyperbole—it is a structural recognition that the early investors in the picks and shovels are now facing a liquidity trap. The capital is there, but the yield is not.
Core: The Data Behind the Rotation
Let me break down the numbers. Nvidia's data center revenue grew over 100% year-over-year in fiscal 2024, but the stock's price-to-earnings ratio during that period surged to over 70x, far above the semiconductor industry's historical average of 15-20x. The hyperscalers, meanwhile, are spending billions on AI-specific chips and data centers, yet their AI revenue contribution remains in the single digits as a percentage of total revenue. This is a classic case of 'pricing in the future' without the present validation.
From my own experience modeling liquidity in DeFi during the 2020 summer, I learned that liquidity depth is an illusion. When capital flows into a single asset class, the exit door narrows. The Chinese hedge funds are not the first to move—I have tracked similar rotation patterns in macro hedge funds over the past six months. The correlation with the US dollar liquidity index is striking: as the dollar strengthens, the appetite for AI infrastructure stocks wanes. The 'super bubble' is not just about AI; it is about the entire macro cycle.
Based on my audit work in 2017, I observed that the most dangerous moment in a bull market is when the narrative becomes self-absorbing. Every conference, every tweet, every analyst report reinforces the same story. The Chinese funds are now challenging that consensus. They are saying that the value creation is shifting from the infrastructure layer to the application layer, and the infrastructure is overpriced. This is not a wholesale exit from AI; it is a rotation into asymmetry.
Contrarian: The Decoupling Thesis and the Crypto Angle
Here is the contrarian view: the rotation out of Nvidia and hyperscalers does not mean AI is dead. It means the market is pricing in a decoupling—the infrastructure story is too crowded, and the real alpha lies in the niches. But the market's reflexivity is dangerous. Fractures in the ledger reveal the truth of value. The capital that leaves the hyperscalers may not flow into application stocks; it may flow into decentralized compute networks, where the cost structure is transparent and the supply is not concentrated in the hands of a few entities.
During my time analyzing the NFT liquidity siphons in 2021, I saw that capital flows from one speculative bubble to another. The Chinese hedge funds may be moving into AI-native crypto projects, such as decentralized GPU networks or AI-driven DeFi protocols. This is the 'crypto countermove'—the idea that the AI infrastructure of the future is not built on centralized cloud providers but on open, permissionless compute. The regulatory environment in Hong Kong, which I have studied closely, is pushing for digital asset licensing, but the real play is in capturing the liquidity that flees from the traditional AI bubble.
Entropy is the only constant in liquid markets. The Chinese rotation is a signal that the current AI infrastructure market is at a point of maximum entropy—the order of the narrative is breaking down. The next phase will be chaotic, but it will also present opportunities for those who understand the technical underpinnings of both AI and blockchain.
Takeaway: Positioning for the Cycle
The takeaway is not that AI is a bubble and you should sell everything. It is that the cycle is turning. The capital that is leaving the hyperscalers will find new homes: application-layer AI, decentralized compute, and even the crypto-native AI sector. The key is to watch the liquidity flows, not the headlines. Over the next 12 months, we will see a reallocation of capital that will create new winners and losers. The question is not whether the AI bubble will pop—it is whether you are positioned for the rotation.
Entropy is the only constant in liquid markets. Fractures in the ledger reveal the truth of value. The truth here is that the infrastructure is overpriced, and the real value is being built elsewhere. The Chinese hedge funds are just the first to see the fault lines.