Avalanche CEO Warns AI Agents May Strain L1 Block Space

AI agents could soon put real pressure on Layer 1 block space, Avalanche Treasury Co. CEO Bart Smith told The Block.

Why Block Space Demand Is Set to Rise

Traditional markets are already moving toward longer trading hours while automated strategies shift onchain. That combination means more transactions competing for the same limited capacity on networks like Avalanche. It's not just about speed anymore; it's about whether the chain can absorb the volume without fees spiking or delays stacking up.

Smith's point lands because AI agents don't sleep. They'll execute trades, rebalance positions and respond to signals around the clock. Once those agents start handling larger notional sizes, the throughput requirements jump fast. We've seen bursts of activity during volatility events before. This could turn those bursts into a steady state.

The Timing Behind the Warning

Expect volatility. Crypto markets have spent years scaling for retail and institutional flows, yet the next wave looks different. It's driven by code that can act without human input and that scales instantly. Avalanche isn't alone in facing the question, but its focus on subnets and custom chains gives it one path to handle the load without congesting the mainnet.

Recent moves by TradFi platforms to extend session times only add to the case. When equity and derivatives venues trade closer to 24 hours, the overlap with crypto never closes. AI agents sitting in both worlds would route orders wherever execution is cheapest and fastest. That arbitrage itself consumes block space.

Market Positioning and Capacity Concerns

Traders already watch gas metrics on Ethereum and equivalent metrics elsewhere. If AI volume materializes the way Smith suggests, those dashboards will need to track agent-driven activity separately. It's one thing to see a DeFi protocol spike; it's another when thousands of independent agents are constantly scanning and acting. The numbers aren't here yet, but the infrastructure bets are being placed now.

Anyone running high-frequency strategies on Avalanche or similar L1s will feel the change first. Fees could become more variable, and latency-sensitive flows might migrate to chains that expand capacity quickest. There's no guarantee the growth arrives overnight, yet the direction is clear once automation crosses from experimental to production.

Watch how Avalanche and its competitors allocate resources between general-purpose execution and specialized subnets. The latter could absorb some of the agent traffic without crowding out other users. Still, every chain faces the same underlying math: finite slots per second against potentially unbounded demand.