The total value locked within decentralized finance (DeFi) platforms hit a record high of nearly $260 billion in December, underscoring sharp investor interest. Yet as regulators in the U.S. and beyond look to control this fast-growing marketplace, DeFi projects now face an increasingly perplexing “decentralization dilemma.”

On one side are users, who have a strong preference for services they control, often in the form of a distributed autonomous organization (DAO), which operates via transparent encoded rules controlled by the organization’s members rather than a centralized party. This is for several reasons. First, users know user-owned services will evolve to benefit users over time, rather than seek to maximize profit. Second, users want to be sure that no single party can take user funds that are locked within the product.

DeFi DAOs achieve user-based governance by distributing governance units, commonly known as tokens, to users. Unlike other types of crypto tokens that are expressly created to be digital currency, governance tokens bestow the right to vote or another form of decision making power. Users often demand these tokens because in many cases they are transferable and can accrue substantial value on the open market. These preferences make it difficult for a centralized company to compete with a decentralized project when it comes to user desirability.

On the other side of this dilemma are regulators. U.S. securities law implies that because governance tokens offer potential for appreciation and are initially issued by a DAO’s founders, they may be securities. As securities, they would face much stricter regulation, including identity checks of token holders, which would undermine the user experience.

In 2018, during the ICO boom, James J. Park, a UCLA law professor, dubbed a narrower form of this conundrum “the Hinman paradox.”

“A token can only be widely distributed to the public if the project it is associated with is functional,” Park wrote in 2018. “But a blockchain project can only be functional if its tokens are widely distributed.”

Regulatory guidance has since begun to suggest that functionality is not enough; a token can only be widely distributed to the public if the project it is associated with is also widely used and governed by those users rather than by a centralized team. A token issued prior to user-based governance, therefore, is not a token but a security.

This presents a classic Catch-22: Governance tokens can only be issued if they are already widely used to make decisions, which would in turn render their issuance unnecessary. It doesn’t take a professor of logic to see the M.C. Escher-esque quandary here. How is an entrepreneur to proceed? This chicken-and-egg conundrum is the heart of the decentralization dilemma.

It’s crucial to note that the dilemma only applies to companies subject to U.S. regulations. Securities laws in other countries can take a very different view. The result is that U.S. companies are at a competitive disadvantage, one that is potentially existential given the essential role governance tokens play in establishing a sizable, distributed user base.

Founders often blame U.S. regulators for this problem, but they don’t have an easy task. Set aside the many tokens that are indeed securities, today’s regulators are mandated to apply decades-old securities laws written with the implicit assumption that every service is overseen by a specific, legally liable entity.