Max Bareiss, head of lending at Galaxy Digital, thinks the next wave of trading may not need a human trader at the desk. In a November 9, 2026 conversation, he argued that decentralized finance is a natural home for agentic trading, where software agents borrow, trade, short, and repay onchain without a person approving every step.
The case is fairly direct. DeFi lending rules are written into the protocol. Liquidation thresholds are set before a position opens, and liquidity sits onchain for anyone who meets the collateral terms. A traditional prime brokerage relationship usually includes a human somewhere in the chain, approving terms or making margin calls. For an AI agent that runs all day, that human is a bottleneck.
Galaxy has already tested the model
Bareiss is not talking only about a possible future. Galaxy has used Aave since at least September 2025, borrowing stablecoins against large collateral positions such as Bitcoin and Ethereum. The setup now supports treasury, trading, and structured product work. The main draw is access to liquidity at any hour, with less dependence on traditional intermediaries.
That timing fits a busy period for the firm. Galaxy announced a $100 million allocation into the Sky project around November 6, 2026. A day earlier, on November 5, Galaxy Onchain Financing Rate, or GOFR, began offering institutional access to onchain capital markets. GOFR pulls together borrowing rates from platforms including Aave, Morpho, and Spark. Galaxy also offers first-loss capital protection and execution services for institutional clients in onchain money markets.
The risk is in the parameters
For investors, the near-term question is where automated trading activity might cluster. DeFi can offer programmable borrowing and atomic settlement, so a trade and its settlement happen together or not at all. Galaxy is building several pieces at once: a rate benchmark in GOFR, a sizable Sky allocation, and client protection. That may give it an early position if agent-driven trading grows, though it is still early days.
The risks deserve attention. An agent that borrows and trades without human oversight depends on how well its limits are set. Pre-set liquidation thresholds can be useful when markets move normally. They can also speed up trouble if many automated positions hit those thresholds at the same time. I think that is the part institutions will watch closely.
What to watch from here: whether GOFR becomes a reference rate across Aave, Morpho, and Spark, how Galaxy deploys its Sky allocation, and whether other institutional lenders start treating AI agents as clients rather than tools.
