Protecting the people building DeFi infrastructure

Welcome to our institutional newsletter, Crypto Long & Short. This week:

• Jennifer Rosenthal on the need to protect the people actually building DeFi infrastructure.
• Alexis Sirkia on how Ethereum’s L2 strategy is failing due to a fundamental design flaw.
• Top headlines institutions should pay attention to by Francisco Rodrigues.
• Aave’s Market Share Slides After rsETH Exploit in Chart of the Week.

-Alexandra Levis

Expert Insights

Protecting the people building DeFi infrastructure

By Jennifer Rosenthal, chief communications officer, DeFi Education Fund

There has been a consistent uptrend in traditional finance companies announcing DeFi-related initiatives, and it’s exciting that these companies embrace technology innovations that will serve as infrastructure for 21st century finance. There seems to also be a growing understanding that open-source, permissionless, programmable, noncustodial, globally accessible and interoperable technology presents major upgrades for certain parts of the financial system.

If you are new to decentralized finance (DeFi), intend to rely on DeFi or want to connect your customers to DeFi, we at the DeFi Education Fund, a nonpartisan, nonprofit organization, invite you to join us in helping to protect the technology and infrastructure that makes it valuable. There are some high-level policy objectives we believe worth defending:

1. Protecting Software Developers and Infrastructure
2. Preserving Self-Custody
3. Advocating for Open Access and Interoperability
4. Championing Permissionless Blockchain Infrastructure and DeFi Markets
5. Supporting Clear Laws and Policies

For months, my team has participated in productive bipartisan, bicameral discussions with members of Congress. We have been impressed by how many Congressional leaders have engaged productively and in good faith to build legislation that reflects a fundamental understanding of neutral, decentralized technology. Software developer protections have come up as a topic of conversation in recent market structure and broader crypto policy discussions. Why? A majority of industry participants agree that if we’re going to use DeFi, we have to protect the people building it.

For example, on February 26, 2026, Representatives Scott Fitzgerald (R-WI), Ben Cline (R-VA) and Zoe Lofgren (D-CA) introduced the bipartisan Promoting Innovation in Blockchain Development Act of 2026 (PIBDA) to protect software developers — who write code but do not control other people’s money — from inappropriate misclassification under criminal code Section 1960. PIBDA clarifies that Section 1960 applies only to those that control customer assets and transmit funds on behalf of customers, aligning the statute with congressional intent and the Treasury Department’s long-standing regulatory interpretation.

In discussing the bill, Rep. Scott Fitzgerald (WI-05) said: “For years, innovators and software developers have been caught in the crosshairs of an aggressive regulatory approach that treats them like criminals. The Promoting Innovation in Blockchain Development Act draws a clear line between those who develop and deploy blockchain software and those who actually move or manage funds. It provides long-overdue legal clarity, protects innovation here at home and allows law enforcement to focus on genuine criminal activity rather than chilling American technological leadership.”

Like the early internet in the 1990s, blockchain technology is a novel innovation evolving faster than existing regulation. Engineers developing open, disintermediated systems do not neatly fit into financial regulations designed for a system that assumes the existence of intermediaries.

As more individuals and companies interact with decentralized infrastructure, our shared voice can play a constructive role in shaping thoughtful and durable policy outcomes. We should collectively support legislative and regulatory initiatives that foster clarity, reduce uncertainty and enable responsible participation across both centralized and decentralized markets.

Thank you for taking DeFi’s tools and technology seriously, and I hope you will join us in defending the policy principles that make building and using DeFi possible.

Principled Perspectives

Ethereum’s scaling problem was never about throughput

By Alexis Sirkia, chairman and co-founder, Yellow Network

Vitalik Buterin recently conceded that most Layer 2 networks are fragmenting Ethereum rather than scaling it. He’s right, but the diagnosis doesn’t go deep enough. The rollup model was never going to deliver a unified scale because it was designed around the wrong assumption: that Ethereum’s limitation was throughput, when the actual constraint was always how value moves between participants.

Rollups addressed congestion by creating parallel execution environments, each processing transactions independently and posting compressed proofs back to the base layer. On paper, that increases capacity. In practice, it produced dozens of isolated liquidity pools that can’t interact without routing assets through bridge infrastructure. The concentration is stark: Base and Arbitrum now capture 77% of all L2 decentralized finance (DeFi) total value locked (TVL), while usage across smaller rollups has declined 61% since June 2025. The long tail is collapsing, and the capital that remains is fragmenting further. Bridge infrastructure has bled $2.5 billion since 2021 for a simple reason: every time value moves between rollups, it passes through a custodial chokepoint. Attackers don’t need to break the chains on either side, they just need to compromise what sits in between.

The industry responded to each bridge exploit by building better bridges. That instinct, while logical at the time, was wrong. The vulnerability isn’t in the bridge implementation. It’s in the premise that value needs to pass through an intermediary at all. State channels eliminate that premise entirely by allowing participants to transact peer-to-peer off-chain, with the base layer serving as the enforcement mechanism rather than the transaction processor. Settlement touches the blockchain only once state-channel transacting finishes, and either party can invoke on-chain enforcement at any point if the counterparty misbehaves.

This isn’t an incremental improvement on the rollup model, but rather a rejection of the assumption that created the fragmentation in the first place. Where rollups multiply execution environments and then try to reconnect them, state channels keep participants connected from the start and only engage the base layer when finality is needed.

The CFTC is preparing to approve the first U.S. framework for perpetual futures, which will pull a meaningful share of $14 trillion in offshore derivatives volume into regulated venues. To put the scale of that shift in context, U.S.-regulated platforms currently handle just 1.6% of global crypto derivatives volume. The infrastructure that absorbs even a fraction of the remaining 98.4% needs to settle cross-chain, in real time, without passing through custodial chokepoints. Rollups, by design, are not candidates for the job.

The 21Shares prediction that most L2s won’t survive 2026 feels pessimistic, but the reason matters more than the timeline. Rollups failed to deliver a unified scale because they treated Ethereum’s constraint as a throughput problem. The market is starting to price in that the real constraint was always trust at the intermediary layer, and the infrastructure that eliminates that layer entirely is where capital and builders will migrate.

Headlines of the Week

By Francisco Rodrigues

This week’s headlines highlight that while the bridges between traditional finance and the crypto sector keep on growing, the devastation caused by smart contract exploits is hitting the market.

Chart of the Week

Aave’s Market Share Slides After rsETH Exploit

Aave’s TVL market share has dropped sharply from ~51.5% in February to ~39% today following the April 18 KelpDAO rsETH exploit, which froze rsETH markets and triggered deposit withdrawals. Active loan share proved stickier, falling only ~2% (54% to ~52%), as existing borrowers couldn’t easily unwind. The AAVE token is down ~50% from its January peak, pricing in both bad debt risk and the reputational cost of being DeFi lending’s largest venue when a collateral asset failed.

Listen. Read. Watch. Engage.

Note: The views expressed in this column are those of the author and do not necessarily reflect those of CoinDesk, Inc., CoinDesk Indices or its owners and affiliates.