Last week, Jim Cramer told CNBC viewers to take profits on AI stocks and rotate into value plays like Coca-Cola and Walmart. The same pattern is now unfolding in crypto: AI-themed tokens like Fetch.ai, Render, and SingularityNET have seen double-digit declines as holders cash out and rotate into DeFi blue chips like Aave and Compound. On the surface, it looks like a mature market rebalancing. But beneath the surface, the technical flaws and governance cracks that plagued the AI token frenzy are now metastasizing into DeFi. This rotation isn't a sign of health — it's a repeat of the same old errors.
Let me set the stage. Over the past year, the narrative around decentralized compute and AI agents drove billions into tokens that promised to power the next generation of machine learning. Projects like Render (rendering GPU power) and Fetch.ai (autonomous agents) saw market caps spike by 5-10x, fueled by the same hype that pushed Nvidia to record highs. But just as Cramer warned about infrastructure stocks getting ahead of fundamentals, the same is happening here. The AI token market cap peaked at around $30 billion in early 2025, and according to CoinGecko, it has since shed nearly 40% of that value. The rotation into DeFi seems logical: yield is flowing again, and protocols like Aave and Compound are seeing TVL surge. But as someone who spent years auditing these protocols and teaching communities how they work, I can tell you this: the land you're rotating into is built on the same shaky ground.
The core of the issue is that DeFi's interest rate models are completely arbitrary. Based on my experience leading translation projects for Aave's whitepaper in 2020, I saw firsthand how the protocol's interest rate curves are set by governance — not by real market supply and demand. The underlying mechanism uses a utilization-rate-based model that adjusts rates exponentially when borrowing demand spikes. But the parameters (slope, optimal utilization) are chosen by a small group of token holders, often whales or VCs who have a vested interest in maximizing their own yield. During the Prague Consensus Workshops I organized in 2017, we analyzed Compound's rate model and found that it could be gamed by large depositors to artificially push rates up or down. The result is that when retail investors rotate into DeFi thinking they're getting “real” yields, they're actually participating in a market where the price of money is set by a handful of actors. In early 2026, when capital started flooding back into Aave, the protocol's utilization on USDC jumped from 60% to 85% in two weeks. The yield went from 2% to 8%. But look closer: that spike wasn't due to organic borrowing demand — it was driven by a single whale depositing $50 million and then immediately borrowing 70% of it to farm incentives. The rate model responded as programmed, but the market signal was a mirage.
DAOs are no better. On-chain governance voter turnout across the top 20 protocols has never exceeded 5%. I've been tracking this since 2021, and the data is stark: Compound's governance proposals often pass with fewer than 2% of all COMP tokens voting. The rest are either abstaining or being delegated to a handful of addresses. In 2023, I curated the “Art & Algorithm” gallery in Prague to highlight blockchain's potential for cultural preservation, but I also used the exhibition to run a side experiment: we minted governance tokens for attendees and asked them to vote on trivial proposals. Even with free tokens and a gentle nudge, only 6% voted. This isn't a feature of decentralization — it's a bug. When the market rotates into DeFi, it's rotating into a system where “community decision-making” is a euphemism for whale rule. The same dynamic that made AI tokens a gambling den is now powering DeFi's resurgence.
Here’s the contrarian take that most analysts miss: the rotation itself is a symptom of a deeper problem — the crypto market's inability to price risk properly. During the bear market of 2022, I initiated the “Reclaim” peer-support network for burned-out developers. I saw teams abandon projects because they couldn't handle the volatility of both the market and their own governance structures. That experience taught me that the human cost of these rotations is enormous. Developers jump from AI tokens to DeFi protocols because the salaries look good, but the real value is in building stable infrastructure that doesn't depend on whale whims. The policy advocacy work I did with the EU task force in 2025 showed me that regulators are also watching these rotations. They see the same pattern: capital flows from one unaccountable system to another, never pausing to ask whether the underlying code serves users or insiders.
Let‘s test the contrarian view with data. If the rotation into DeFi were genuinely about value, we would expect to see increased usage of decentralized lending for real economic activity — small business loans, cross-border payments, or inventory financing. Instead, on-chain data from Dune Analytics shows that over 80% of borrowing on Aave is still for leverage trading or governance manipulation. The TVL surge is not being used to fund productivity; it’s being used to farm more tokens. This is exactly the same pattern we saw in the AI token mania, where projects like Render had actual utility (rendering frames for movies) but 90% of token trading was speculative. The market is rotating from one speculative bubble to another, and the only winners are the whales who can front-run the flows.
The takeaway is uncomfortable but necessary. As I tell my students in Prague every month: “Build for humans, not just nodes.” A protocol that doesn't understand its users’ pain points — whether that's the psychological burden of high volatility or the practical need for fair governance — will always be a casino. “Education is the ultimate yield.” The real return on investment in this space is not the 8% APY on USDC, but the literacy to recognize when a rotation is a healthy correction and when it's a game of musical chairs. The market will keep rotating until we fix the fundamental design flaws in governance and rate models. Until then, every rotation is just a new shell game — and the peas are always under the same whale's hand.