When the world’s largest bank tells you to spread your AI bets, it is not optimism—it is a structural admission that the easy alpha has been captured. The ledger bleeds red when trust decays into code, and here the trust is in the singularity of the AI narrative. JPMorgan’s Gabriela Santos has publicly suggested diversifying AI investments across regions and industries. For a macro watcher, this is not a portfolio tip; it is a signal of phase transition in the technology cycle. The money that once flowed into single GPU stocks and cloud giants is now being told to spread thin. That is a symptom of maturity, not opportunity.
Context: Global liquidity maps are redrawing. After the 2023–2024 infrastructure build, AI capital expenditure peaked, and the signal from institutional strategy desks is clear: the next phase belongs to application diffusion and vertical integration. This mirrors what I observed in the crypto space in 2025 when BlackRock’s BUIDL fund integrated with Ethereum Layer 2s. I quantified how tokenized real-world assets reduced settlement times by 94%—yet the narrative remained stuck on the infrastructure layer. Now, AI is undergoing the same transition. The JPMorgan note is not a standalone idea; it is part of a broader convergence where institutional capital begins to treat technology sectors as multi-asset ecosystems rather than single-theme bets.
Core: The diversification thesis rests on a valid but fragile assumption: that AI’s value chain has become sufficiently granular to allow risk dispersion. From my analysis of the FTX collapse, I learned that hidden leverage in concentrated positions creates systemic fragility. The same applies here. AI’s valuation expansion has been extreme—by mid-2025, the forward P/E of AI-exposed stocks was 40% above the market average. Diversification is a hedge against the inevitable reversion. But the key insight is not the strategy itself; it is the timing. When a top-tier institution publicly recommends dispersion, it usually means the concentrated winners have already repriced. In my work decoding the digital euro pilot, I discovered that offline transaction caps of €300 were a design choice that limited utility. Similarly, the AI diversification signal is a design choice to limit downside, not to capture upside. The real value lies in the second-tier assets—the vertical AI players in healthcare, manufacturing, and finance that have not yet been fully priced. We are auditing the ghost in the machine’s soul, and what we see is a capital rotation that rewards specificity over breadth.
Contrarian: The blind spot in this strategy is the assumption that AI’s technological pluralism will persist. I have seen this pattern before in crypto’s Layer 2 narrative. ZK rollup proving costs remain absurdly high; unless gas returns to bull-market levels, operators are bleeding money. The same risk exists in AI: if the model layer converges to a single dominant paradigm—say, a general-purpose AGI from one lab—the diversification benefit evaporates. Cross-regional bets also carry geopolitical risk. When I analyzed the liquidity model for tokenized RWAs, I found that cross-border settlement times improved but compliance costs rose. The decoupling thesis—that AI and crypto can be treated as separate macro assets—is fragile. Both are driven by the same liquidity cycles and the same regulatory mood. The ledger bleeds red when trust decays into code, and trust in AI’s decentralized future is already fraying as compute becomes a national security asset. Diversification in a hyper-correlated environment is an illusion.
Takeaway: The question is not whether to diversify, but whether the diversification itself is a hedge against a system that is already fracturing. We are auditing the ghost in the machine’s soul. As liquidity tightens and the tech cycle matures, the real positioning is not in spreading bets, but in identifying the one chain—or the one vertical—that will survive the convergence. The JPMorgan note is a canary in the macro coal mine. Read it not as advice, but as a confession that the easy money has moved on.