What Brings Capital Back to DeFi? | Roundup

Recent DeFi failures have made security and accountability central to whether capital returns. Drift and Kelp DAO each suffered exploits of roughly $285 million, and the Lazarus Group reportedly obtained about $2 billion from DeFi and crypto hacks last year. Kelp’s rsETH was accepted as collateral on Aave Core, transmitting risk through a shared pool; speakers characterized the episode as a collateral and pool-design problem, not an Aave multisig or oracle failure. Aave deposits fell from about $45 billion to $31 billion in a week, with one later estimate putting them just under $30 billion. Utilization also rose as some users could not move wrapped ETH.

The Kelp incident also exposed risks in cross-chain infrastructure: Kelp operated its LayerZero DVN and configured it as 1-of-1, creating a central point of failure. One speaker estimated that about 97% of LayerZero DVN setups were 1-of-1 or 2-of-2, while the discussion left responsibility between Kelp’s configuration and LayerZero unresolved. Speakers said the industry had favored speed and scalability over security, and that increasingly sophisticated attacks—including infrastructure infiltration—are pushing attention toward protocol and smart-contract defenses. One expected DeFi to recover after bottoming, potentially over a year or two or longer, but whether AI or the anticipated Mythos release explains the apparent acceleration in attacks remained uncertain.

Lower interest rates could help bring capital back by making DeFi yields more competitive with bank deposits and encouraging borrowing and speculation. Under those conditions, speakers said USDC deposit rates might return to roughly 5–6%, but they disagreed on whether rate cuts alone would be enough. With yields around 4–6%, some investors see little reward for taking unclear protocol risks when a bank can offer 5% or Treasuries can offer more than some DeFi opportunities. Speakers said contagion risk and low yields currently make them reluctant to commit substantial capital. Most funds leaving Aave were expected to move off-chain, though Spark Lend nearly doubled deposits and Morpho appeared to be recovering after a decline.

Lending models distribute trust differently rather than removing the need for it. Aave’s structured core differs from more isolated or modular approaches associated with Kamino and Morpho, where vault users rely on curators to select assets. Speakers noted losses tied to asset selection or curator-related risks in both Aave and Morpho. Large pooled models can face withdrawal cascades and bad debt; KYC or AML checks, circuit breakers, and withdrawal limits might reduce some risks, but at the cost of composability. Users also need clarity about who stands behind a vault strategy. More credible curators willing to accept liability could help rebuild confidence, although even well-regarded curators can lose users’ money.

Real-world assets may be prominent in DeFi’s next cycle, as current borrowing and lending activity shows more interest in RWAs than crypto-native assets. Off-chain representation may suit the present risk environment, while on-chain yield-bearing RWAs could enable new savings and looping strategies. Speakers did not agree on which ecosystem will lead: Ethereum projects built on Morpho were described as having an advantage to lose, while Kamino has found success and Solana is also focused on RWAs. Institutional capital and stablecoin distribution could bring new users on-chain, but institutions need confidence in security before making billion-dollar allocations. One speaker expected institutional adoption on roughly a one-year horizon while warning that the next 12 months could be difficult. GENIUS and CLARITY, and encouragement from SEC and CFTC leaders, were cited as support for rebuilding finance on these rails.

Speakers differed on whether agents or fintechs are the likelier adoption route. Agents could deploy more capital over the next few years, selecting backend protocols or APIs for reliability; Morpho had been working for two or three weeks on agents that could execute through its backend. But another speaker argued fintechs are the more immediate opportunity, while the response was that users may prefer 5.5% Treasuries to 3.5% on Aave when DeFi carries loss risk and offers a weaker user experience. Neither route is yet clearly driving adoption. The discussion also pointed to sticky ecosystems: deposits showed little apparent movement between Ethereum and Solana, and users often deploy funds within the ecosystem where they receive USDC. Speakers anticipated more controls and reduced composability in larger-scale DeFi alongside a smaller permissionless niche, with consolidation likely over the next couple of years and security remaining a backend prerequisite for growth.