Over the past seven days, on-chain data revealed a stark rotation: AI-token-related protocols absorbed 60% of all new crypto capital inflows, while top-10 DeFi TVL dropped 15% in the same period. This isn’t a crypto-native anomaly — it’s the same macro force that JPMorgan strategist Fabio Bassi recently identified for traditional markets: AI dominance is creating a capital vortex that leaves entire ecosystems underinvested.
Bassi’s argument is simple: European stocks will continue to underperform globally because the region lacks exposure to the AI revolution. Structural headwinds — high policy rates, elevated energy costs, and chronically low productivity — make it impossible for Europe to participate in the productivity shock that AI represents. In crypto, the analogy is uncomfortable but precise. Just as European equities bleed to US tech, capital flows in digital assets are pivoting away from foundational DeFi and Layer-2 ecosystems toward anything with an AI label.
Let’s ground this in technical reality. Over the last quarter, daily active users on AI-centric crypto protocols (like decentralized compute networks and agent frameworks) grew 220%, while user activity on major DeFi lending pools remained flat or declined. TVL for AI-token projects surged 340%, versus a 12% drop for Aave and Compound. This mirrors the “quality dislocation” Bassi describes — capital isn’t leaving risk, it’s rotating to the only narrative that promises exponential growth. In traditional markets, AI acts as a quasi-monetary policy, creating its own credit cycle. In crypto, AI tokens are doing the same: they command higher yields, attract faster market-making, and enjoy lower friction from regulatory scrutiny because regulators are still learning what “AI agent” means.
But here’s where my experience as a DeFi PM grounds the analysis. In 2020, I led product strategy for a lending protocol and realized the “code is law” ethos was masking centralized oracle manipulations. That same pattern repeats today. The AI projects soaking up capital are often built on questionable foundations — centralized sequencers, opaque tokenomics, and governance that delegates power to KOLs who cannot even read the source code. DeFi’s core promise — permissionless, transparent lending — is being abandoned for high-APY pools that will disappear the moment the AI hype cycle turns. This is exactly the dynamic I saw in DeFi Summer: liquidity mining APY is the project subsidizing TVL numbers. Stop the incentives, and real users vanish. The same applies here.
Bassi’s contrarian angle is that European high-quality exporters might benefit from global resilience, but that benefit is parasitic — it depends on US AI strength. In crypto, the contrarian truth is harsher: the AI gold rush is building on the same fragile assumptions that burned DeFi. Many “AI tokens” have no genuine product, only a ChatGPT wrapper and a whitepaper about decentralized inference. The structural problems that plague Europe — high energy costs, fragmented regulation, and a lack of venture funding for deep tech — also apply to running large models on-chain. Code betrays when we do. When I audited the Zilliqa mainnet in 2017 and discovered a consensus race condition, I learned that speed without integrity creates debt. Today, AI crypto projects are accumulating technical debt by skipping rigorous testing and governance design, betting that narrative will outrun reality.
Burnout is the tax on innovation — and it’s being extracted from developers who are pushed to ship AI agents without thinking about human-centered security. The same spiritual hollowness I felt during the 2021 NFT explosion is returning: we are building digital vanity metrics again, this time with GPUs. But there is a path forward.
The real opportunity lies not in chasing AI tokens, but in building systems that verify human intent in an age of synthetic media. Decentralized identity protocols, zk-proofs for proof-of-personhood, and verifiable computation layers — these are the infrastructure that can survive the narrative cycle. When the AI bubble cools, the capital will remember that trust requires more than a whitepaper. It requires algorithmic empathy: a verifiable layer of human accountability. That vision is why I argue for human-centric decentralization — so that as AI grows, our decentralized structures remain rooted in values, not just token prices.
The macro lesson from JPMorgan is clear: structural advantages compound. Europe’s lack of AI productivity will keep it lagging. Crypto’s equivalent — a lack of real verifiable utility — will leave many AI projects stranded. But for those who build with patience and integrity, the winter is the best time to plant seeds.
