TL;DR:
- SharpLink CEO Joseph Chalom projects AI agents could save investors $1.4 trillion annually by 2035 by intensifying fee competition across global finance.
- His model sees more than $1 trillion in financial-services revenue becoming contestable by 2030, rising to $4 trillion by 2035.
- Chalom sees stablecoins and Ethereum as potential rails for agentic finance, though closed fintech ecosystems could also capture activity as companies compete for control and customer financial relationships.
SharpLink CEO Joseph Chalom says AI agents could cut financial-services fees by $1.4 trillion annually by 2035 as autonomous software reshapes how consumers save, invest and pay. In a public post, Chalom said his team modeled ten financial verticals and expects more than $1 trillion of industry revenue to become contestable by 2030, rising to $4 trillion by 2035. The forecast assumes agents force providers to compete harder on price, potentially returning $350 billion to users annually by 2030 and $1.4 trillion five years later.
— Joseph Chalom (@joechalom) September 23, 2026
AI Agents Could Reshape Financial Fees
Chalom also points to $15 trillion held by U.S. households in checking, savings and short-term deposit accounts, arguing that much of that money earns below money-market rates. He estimates savers forgo at least $180 billion annually because cash remains parked in lower-yielding products. AI agents could automatically compare options, move idle balances and negotiate financial services on users’ behalf. That vision overlaps with emerging AI-agent payment systems designed to let software transact without constant human intervention.
The competition is drawing banks, brokers, payment companies and crypto firms. Chalom named Visa, Mastercard, Stripe, PayPal, Circle, Tether, Robinhood, Coinbase and Binance among companies positioning for the agent economy. Whoever controls the financial infrastructure could influence which products agents recommend and where customer funds move. Stablecoins are emerging as a natural payment rail because autonomous software needs programmable, always-on settlement, a use case already visible as AI agents begin operating stablecoin wallets and executing machine-to-machine payments.
Chalom believes much of that activity could settle where stablecoins, tokenized assets and DeFi liquidity are already concentrated, placing Ethereum in a strong position. SharpLink itself held 891,714 ETH as of September 14, linking the company’s treasury strategy to that broader thesis. The argument also aligns with the push to build programmable payment rails for autonomous software, where open blockchain infrastructure competes with closed fintech systems.
The outcome is far from settled. Fidelity has argued that technology and fintech companies could capture agent activity inside closed ecosystems instead of public blockchain payment standards. For investors and financial firms, the larger question is whether AI agents primarily compress fees, redirect customer relationships or shift financial activity toward new settlement networks. Chalom’s model suggests all three could happen at once, turning automated finance into a battle over distribution, infrastructure and trillions in annual revenue.
