🧠 Food for thought

Implications, context, and why it matters.

No evidence autonomous agents work safely yet

• Donut has US$15 million in funding and a 160,000-person waitlist, with no live beta results, backtesting data, or third-party audits.
• It raised US$7 million less than six months ago and the product is still in development. The team claims autonomous market analysis and onchain execution. Here, autonomous agents means software that makes and executes trading decisions without human approval.
• The team has not named which DEXs (Decentralized Exchanges), chains, or execution layers (where transactions are processed) it uses. That gap raises questions on slippage, MEV (Maximal Extractable Value) attacks, or liquidity during onchain execution.

MEV protection infrastructure creates partnership opportunities

• Trading front-ends like Donut create openings for RPC (Remote Procedure Call) providers with MEV-protected endpoints. GetBlock (an RPC node provider) integrates with Merkle’s private mempools (non-public queues for pending transactions) 1. It supports Solana and Ethereum. It adds support for BNB Smart Chain (BSC) and Base (an Ethereum layer-2 network).
• Blockchain infrastructure startups can target wallets (crypto apps that hold and send assets) and DEXs with private orderflow routing. DRPC (an RPC provider) offers MEV protection on Ethereum and Base 2. It also covers Polygon and Solana 2. Helius (a Solana infrastructure provider) gives Solana-native low-latency access 3. That mix patches holes in execution security for AI trading browsers.
• Compliance and risk tools lag. Third parties that build transaction simulation, guardrails, or audit trails for autonomous agents can win business from Donut-like startups. They can serve institutional users in DeFi (Decentralized Finance) derivatives.