Autonomous AI agents have the potential to become a new catalyst for activity in the digital assets space, spanning a spectrum from payments to complex credit operations. However, as my research shows, the distribution of benefits from this technological shift will be extremely uneven, and not all market participants will end up among the beneficiaries.
Development Productivity: AI as an Accelerator
Analysis of data from more than 100,000 developers on GitHub demonstrates an impressive effect from the adoption of digital assistants. The use of AI tools for writing code increased the number of commits by up to 180%, and the frequency of releases by 30%. This opens doors for small teams, allowing them to build and launch blockchain applications faster. Nevertheless, in critical areas such as finance, manual code verification will remain mandatory—automation will not replace human oversight of security.
It is telling that similar dynamics are observed within the crypto industry itself. In 2026, despite falling prices and a reduction in the number of developers, the average number of changes per specialist continued to grow. This suggests that AI increases labor efficiency even under market contraction conditions. However, it is important to understand: an increase in the number of applications does not guarantee their success. The user base, liquidity, and trust remain key factors.
On-Chain Transactions: A New Era of Autonomy
A real breakthrough is expected from AI agents capable of independently making payments, trading assets, and managing liquidity. Blockchains are ideally suited for this purpose thanks to round-the-clock operation, programmable settlements, and support for microtransactions. The market is already beginning to take shape: according to my data, by May, AI agents had conducted over 176 million transactions worth more than $73 million, with 98.6% of operations involving the USDC stablecoin.
Infrastructure is actively developing. For example, Coinbase is implementing the x402 protocol and a set of tools called Coinbase for Agents. However, I do not expect all agent operations to move to public blockchains. Traditional banks and fintech companies, possessing client bases and credit infrastructure, will develop their own solutions. Most likely, AI agents will become multiplatform, choosing routes based on cost and convenience.
Revenue: Not All Transactions Are Equally Valuable
A critically important point: millions of new transactions do not guarantee proportional growth in network revenues. Payments, although numerous, generate low fees per operation. They can be bundled or moved to cheap L2 solutions. Much more significant value is generated by capital-intensive operations. My calculations show that trading generates 49 times more revenue for Ethereum's base layer per dollar of volume than payments, and MEV creates additional income.
This is why I see the greatest potential in AI agents focused on trading, lending, and providing liquidity. If automatic payments become widespread, the main beneficiaries will be stablecoin issuers and infrastructure providers, rather than holders of native blockchain tokens.
My conclusion: the market overestimates the impact of AI on the value of base protocols. True value will shift toward applications and services that can effectively monetize capital-intensive operations, rather than simple transactions.