In March last year, a DeFi protocol I had audited a month earlier saw its native token price triple on the back of a promised 300% staking yield. I rechecked the on-chain revenue data. The protocol’s actual fees that month covered less than 0.3% of that yield. The rest was printed—new tokens, diluting existing holders, inflating the illusion of profit. Six months later, the token crashed 90%. The yield had vanished, but the loss was real.
We didn’t learn the lesson. Right now, the crypto market is replaying the same pattern that Wall Street strategists warned about in their ‘earnings bubble’ analysis: a concentrated, unsustainable profit expectation propped up by narrative and liquidity, not fundamentals. Only in DeFi, the bubble isn’t about corporate earnings—it’s about protocol revenue and yield promises.
Context: The yield narrative is the new earnings guidance
Wall Street’s current dilemma—stock analysts predicting 25% profit growth while interest rates rise—has a direct mirror in DeFi. Today, the average yield on top lending protocols hovers around 4-6% from real borrowing demand, yet many new ‘flywheel’ projects boast double-digit APYs from token emissions. These emissions are the crypto equivalent of the earnings guidance boost that strategists call a bubble. They are non-sustainable. Based on my experience auditing tokenomic models in 2017, I recognize the signs: the promised yield depends entirely on new capital entering the system, not on underlying economic value creation.
Between January and April this year, total value locked (TVL) across the top five yield-generating protocols grew by 25%, but the projected annualized token inflation to sustain those yields grew by 60%. The divergence is a flashing red beacon.
Core analysis: The structural unsustainability of yield inflation
Let’s dissect one representative example: a ‘restaking’ protocol that launched in December 2024. It promised a 35% APY for staking an ETH derivative. The revenue sources: 0.5% fee on each restaking transaction and a share of sequencer fees from one L2. In March 2025, the protocol’s weekly on-chain revenue averaged $240,000. To pay the advertised yield, it needed to distribute $1.2 million per week in token rewards—a 5x gap. The difference was generated by minting new protocol tokens, which were then sold by farmers, putting downward pressure on the token price. The yield was real in token count, but value was being destroyed systematically.
This is identical to the ‘earnings bubble’ logic identified by analysts like Ben Inker: growth that is not backed by cash flows, only by optimistic assumptions about future adoption. In crypto, the assumption is that TVL will keep growing, so token price will appreciate, offsetting dilution. But as Michel Lerner pointed out about AI stocks—‘profit forecasts assume technology will generate supernormal returns forever’—in DeFi, yield forecasts assume new users will enter forever. Both are fragile.
From my 2020 DeFi community building work, I know that once users feel the inflation, they leave. In a 2023 survey I conducted among 15 DeFi projects, average user retention after a token reward halving was only 12% after three months. The bubble pops when the next new narrative (e.g., ‘AI agents on Bitcoin’ ) pulls liquidity away faster than the protocol can print.
Let me be precise about the numbers. On Ethereum L2s, which host most yield activities, daily transaction fees have grown from $0.85 million in January 2025 to $1.2 million currently—a 41% increase. But the total token rewards emitted by the top 10 yield farming protocols has grown from $45 million to $92 million per month—a 104% increase. The ratio of revenue to reward emission has worsened from 19% to 13%. We are subsidizing growth with dilution at an accelerating rate.
Contrarian angle: Why this bubble is more dangerous than 2021
You might argue that 2021 was worse because liquidity mining APYs hit 1,000%+. True, but back then, the industry was younger, and few expected sustainability. Today, we have maturity. We have audited code, institutional custody, and real yield from DEX fees. The problem is that these ‘real’ yields are now being used as a fig leaf for far larger inflationary promises. The narrative ‘we have revenue’ covers up the fact that revenue covers less than 15% of promised yield. It’s a higher-quality fraud, which makes it harder to detect until the dilution compounds.
Also, the current market structure is more interconnected. Many yield-bearing tokens are used as collateral in lending protocols. A single large protocol’s yield collapse could trigger liquidation cascades across DeFi. In 2021, the Terra collapse was isolated (though severe). Today, a similar event could spread via restaking, cross-chain bridges, and derivative products. The bubble is not just in one protocol—it’s systemically embedded.
Moreover, the ‘earnings bubble’ warning from traditional finance reminds us that when expectations peak, the correction is sharp. In crypto, expectations are constantly peaking because we live on hype cycles. The current bull cycle driven by AI-crypto narrative (autonomous agents, decentralized compute) has inflated not just token prices but also the promised yields of projects riding that wave. If the AI hype falters—say, a major protocol’s agent fails or a regulatory action on AI blockchains—the yield bubble will pop along with it.
Takeaway: The bridge between hope and reality
The Wall Street strategists ended their warning with a question: ‘What happens when growth expectations meet an economy that cannot sustain them?’ We must ask the same in DeFi: What happens when yield expectations meet on-chain revenue that cannot sustain them? The answer is either a painful correction or a slow bleed that erodes trust in the entire sector. The ethical mission we champion—decentralization, transparency, empowerment—depends on building protocols that can survive reality. We didn’t design this bubble, but we can choose to step away from it. Code is law, but empathy is the constitution—and empathy demands that we stop promising what cannot be delivered.