Automated AI agents could turn blockchain capacity from a largely theoretical concern into a practical constraint for financial markets, according to Avalanche Treasury Co. Chief Executive Officer Bart Smith, who said rising machine-driven activity may push Layer 1 networks closer to their throughput limits.
Speaking during a Wednesday interview at the Avalanche Summit in New York, Smith said even conservative assumptions about “agentic” trading and automated financial workflows point to substantially more activity on public blockchains. Unlike human traders, automated programs can operate continuously, submitting transactions, updating positions and responding to market data without pauses between trading sessions.
That pattern would place greater pressure on blockspace, the limited transaction-processing capacity available on a blockchain. When demand exceeds available capacity, users typically compete through higher fees or wait longer for transactions to be included. The result can affect everything from settlement and collateral movements to decentralized exchange activity and stablecoin transfers.
Smith’s argument places the coming debate over financial-market automation squarely on the underlying infrastructure. Networks may face a more demanding workload if AI systems begin handling routine market-making, treasury management, rebalancing, compliance checks and payment flows onchain at scale.
Layer 1 design could become more consequential
Smith said the prospect of persistent machine activity could make architectural differences between major Layer 1 networks more relevant. He named Avalanche, Solana and Ethereum, each of which takes a different approach to transaction processing, validator design, network fees and application deployment.
In an environment with excess capacity, businesses and developers can often choose networks based on liquidity, ecosystem reach or familiarity without treating transaction throughput as a binding limit. Smith said that calculation changes when blockspace becomes scarcer and systems must handle frequent, time-sensitive requests.
The practical question for firms would extend beyond headline transactions-per-second figures. Financial applications need predictable execution, resilience during demand spikes, security controls and, in some cases, privacy features that prevent commercially sensitive activity from being exposed on a public ledger.
Smith pointed to privacy and security as central considerations for businesses assessing blockchain infrastructure. Financial institutions may be reluctant to place certain workflows fully on transparent networks if transaction patterns, counterparties or treasury movements can be easily observed. That leaves room for designs that combine public settlement with more controlled execution environments or application-specific chains.
Avalanche has emphasized its ability to support customized networks, while Ethereum’s ecosystem has increasingly relied on Layer 2 networks for lower-cost activity. Solana, meanwhile, has focused on processing large volumes of transactions on a single high-performance base layer. Smith’s comments suggest capacity planning may become a more immediate commercial issue if automated systems begin generating activity at the pace proponents expect.
A 24/5 market would require different rails
Smith also said he expects traditional financial markets to move toward 24-hour-a-day, five-day-a-week trading by mid-2027. Such a schedule would sharply reduce the operational gap between conventional securities markets and cryptocurrency markets, where transfers and trading are already available around the clock.
A move to 24/5 trading would require more than extending exchange opening hours. Clearing, collateral management, market surveillance, custody, financing and payment systems would need to operate with far less downtime. Smith argued that existing financial infrastructure would struggle to support a broad shift to near-continuous market access without systems built on blockchains.
The appeal of blockchain-based rails is not simply that they run continuously. They can also coordinate asset ownership, payment and settlement within shared systems, reducing the need for multiple parties to reconcile records after trades are executed. That model could shorten settlement cycles and allow collateral to move more frequently, though it also shifts more operational pressure onto the networks processing those transactions.
Automated agents could amplify that pressure. A human-driven market has natural pauses: traders sleep, offices close and manual review slows decision-making. Software agents can monitor prices, hedge exposures, move liquidity or execute pre-programmed instructions every second of the day. Many of those actions may never require a final onchain transaction, but workflows designed to settle, verify or coordinate activity on public networks would add to demand for blockspace.
Circle’s Arc test adds a data point
Circle on September 16 introduced Arc, a network designed for machine-to-machine financial activity, according to the company’s announcement. Circle said Arc processed more than 700 million basic operations during testing and would use USDC to pay network fees.
The project offers an early example of how stablecoin issuers are preparing infrastructure for automated payments and agent-led workflows. Using USDC for fees could simplify payment handling for software agents, which would otherwise need to hold and manage a separate volatile token solely to pay for network access.
Circle said the USDC supply stood above $74 billion worldwide. That scale gives the stablecoin an established pool of onchain liquidity, although large-scale agent settlement remains an emerging use case rather than a proven replacement for conventional market infrastructure.
The comparison also underlines a distinction likely to matter as these systems develop: high operation counts do not automatically translate into the same load as fully settled public blockchain transactions. Networks can batch, compress or process certain actions offchain before recording final results on a ledger. The eventual strain on Layer 1 capacity will depend on how often agents need to settle independently, how applications structure their transactions and whether activity is routed through separate execution layers.
Treasury exposure sharpens Avalanche’s interest
Smith’s position gives his comments an additional corporate dimension. Before joining Avalanche Treasury Co., he served as chief executive of Susquehanna Crypto and spent nearly 14 years at Susquehanna, bringing market-structure experience to a company tied closely to the Avalanche ecosystem.
The supplied material states that public records show Avalanche Treasury Co. holds about 15 million AVAX tokens, equal to roughly 3.5% of the available supply. Such a position gives the firm a direct interest in the long-term demand for Avalanche blockspace and the network’s ability to compete for institutional and automated financial workloads.
Smith’s capacity warning should therefore be read as both a market-structure view and a case for networks designed to accommodate specialized financial activity. The test for Layer 1 platforms will be less about whether automated agents arrive than whether their transactions can be handled reliably when market conditions are busiest and the cost of delayed settlement is highest.
Explore how AI in banking reshapes automation, infrastructure demands, and real-time financial market behavior onchain.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
