PodBrowser
Crypto 101

Ep. 749 Coinbase’s Head of AI Explains the Future of Agentic Finance

Thursday, 3 September 2026 · 3 min read · Listen to the episode ↗

In this episode, Lincoln Mer, Coinbase’s Head of AI, delves into the transformative role of AI in cryptocurrency markets, highlighting the innovative X402 protocol that enables AI agents to perform on-chain activities without incurring Ethereum gas fees. He discusses the rise of AI-powered financial advisors and the potential for AI to lower costs in ETFs and mutual funds, while also addressing the challenges of integrating AI with existing financial regulations.

Lincoln Mer discusses the increasing integration of AI in the cryptocurrency markets, emphasizing its role in analyzing market trends and managing investment portfolios. He introduces the X402 protocol, which allows AI agents to access crypto wallets and perform on-chain activities, such as executing swaps and deploying smart contracts, although its initial development was primarily experimental.

The X402 protocol aims to simplify agent payments for online services by enabling users to allocate limited funds for transactions without the need for Ethereum gas fees. This innovation is particularly relevant as microtransactions gain popularity, especially with stablecoins, given that traditional payment networks often impose high minimum transaction sizes and fees.

Mer highlights that since the launch of the agent economy, over a hundred million transactions have been recorded, with AI agents emerging as the primary consumers of the web. However, a significant 76 percent of these transactions fall below the 30 cent minimum set by card networks, revealing a disconnect with legacy financial systems that are not optimized for AI agents.

Coinbase has introduced an AI-powered financial advisor as part of its Next Bets program, which aims to democratize access to quality financial advice. Currently, this advisor is available to only about 1% of Coinbase One members and can assist with tax loss harvesting and portfolio optimization.

The conversation also explores the potential for AI to reduce fees in ETFs and mutual funds, contributing to a deflationary environment by lowering the costs of products and services. However, existing regulations necessitate human intervention for individual trading actions, which constrains the full capabilities of AI in trading. Despite these limitations, Coinbase's MCP product allows agents to implement autonomous trading strategies, making it easier for individuals without technical expertise to participate.

Mer elaborates on the potential of the X402 protocol to create a two-sided marketplace for data providers and agents, integrating major data sources like CoinMarketCap and CoinGecko. There is a significant opportunity to develop a data marketplace using X402, which supports payments in various cryptocurrencies, with USDC being the predominant currency for agentic transactions.

Setting up agents is more straightforward than anticipated, with users able to establish their agents through specific Coinbase documentation. However, challenges remain regarding identity and tax responsibilities related to agents, as the human owner is accountable for tax payments. Users can also implement manual approvals for agent interactions to mitigate the risk of unwanted transactions.

Looking ahead, predictions suggest that agents will be utilized more effectively in five years, although the specifics remain uncertain. The demand for agentic trading and payments is expected to grow as data becomes increasingly valuable, with the emergence of fully decentralized autonomous agents managing their own resources viewed as an inevitable development.

Lincoln emphasizes Coinbase's ambition to serve as the backbone of the agentic economy through AI finance, which will encompass both agentic trading and payments. While X402 operates independently from Coinbase, it relies on the company's infrastructure for payment processing and aims to establish an open standard to capture value. This approach is compared to Google's strategy with Android, contrasting with Apple's iOS model. User trust and technological advancements are deemed essential for the successful adoption of self-driving-like AI systems in finance.

This summary was generated from the episode transcript and can contain mistakes.