AI Agents Could Slash $1.4 Trillion in Finance Fees by 2035, Says Sharplink CEO

Joseph Chalom wants you to think about what $1.4 trillion actually looks like. The Sharplink CEO put out a forecast recently arguing that AI agents will gut global finance fees at a scale most people haven’t seriously priced in yet — saving investors that full $1.4 trillion per year by 2035. It’s a big number. Probably too big to absorb in one read.

Chalom’s team built a model covering ten financial verticals and ran it out through 2035. The core finding: over $1 trillion in financial services revenue gets contested by 2030. By 2035, that contested figure balloons to $4 trillion. The mechanism is pretty straightforward — AI agents force providers to compete harder on fees, which squeezes margins and passes savings down to consumers. Chalom puts the consumer savings number at $350 billion annually by 2030, scaling up to $1.4 trillion by 2035. He also flags something that tends to get buried in these conversations: U.S. households are currently sitting on $15 trillion in checking accounts, savings accounts, and short-term deposits. Most of that money earns below money-market rates. Chalom’s estimate is that savers are losing roughly $180 billion a year because of that gap. AI agents, in his view, could route that idle cash more aggressively toward better-yielding products.

Not everyone’s convinced the benefits will flow freely.

Who Controls the Agents Controls the Customer

The race is already on. Visa, Mastercard, Stripe, PayPal, Circle, Tether, Robinhood, Coinbase, and Binance are all pushing for position in the AI agent infrastructure space. Chalom’s argument is blunt: whoever controls the agents controls which products get recommended and how idle cash gets managed. That’s not a technology point — it’s a distribution point. Owning the infrastructure means owning the customer relationship, and that’s the real prize here.

A BlackRock research paper backs part of Chalom’s framing. The paper put forward the idea that stablecoins could become the default currency for agent-to-agent transactions. Chalom sees most of that activity concentrating on Ethereum. He cited 3.6 million daily transactions recorded on the network in April as evidence that Ethereum already has the throughput and activity base to function as the primary hub for AI-driven financial flows. Stablecoin adoption across major blockchain networks has grown sharply in recent years, and the idea that agents would transact natively in digital dollars on a programmable settlement layer isn’t a stretch — it’s kind of where the infrastructure is already pointing.

Sharplink itself has skin in this game. The company held 891,714 ETH as of mid-September. That’s a substantial position, and it’s hard to read Chalom’s Ethereum optimism as purely analytical when the firm is sitting on that kind of exposure. Worth keeping in mind.

See also: BlackRock Report Reveals Stablecoins Key to a $4 Trillion AI Economy

Fidelity’s Analyst Sees a Closed-System Risk

Fidelity Digital Assets isn’t buying the fully open version of this story. Analyst Max Wadington put out a more cautious read, warning that tech and fintech firms could absorb much of the agent activity through closed systems rather than open, decentralized networks. That’s a real tension. The optimistic scenario has AI agents routing freely across open rails, driving fee competition and benefiting consumers. The pessimistic one has Google, Apple, or a major fintech locking agents inside proprietary ecosystems, capturing the value themselves and leaving the decentralized vision mostly theoretical.

Wadington’s concern isn’t fringe. Closed systems have won before in tech. App stores, payment networks, social platforms — the default outcome in platform competition tends toward consolidation, not openness. If that pattern holds here, the $1.4 trillion in consumer savings might end up looking more like $1.4 trillion in platform profits.

Chalom clearly disagrees. He keeps coming back to infrastructure ownership as the central variable, and his bet — both intellectually and financially, given Sharplink’s ETH position — is that Ethereum’s open, high-throughput network is where the action concentrates. Whether that’s right probably depends on regulatory choices that haven’t been made yet, and on whether the big incumbents can move fast enough to wall off their ecosystems before open-network adoption gets too deep to reverse.

Related: BlackRock Predicts AI Will Drive Demand for Ethereum and Stablecoins in Finance

The stablecoin angle matters here too. If BlackRock is right that stablecoins become the native currency for agent transactions, then whoever issues and controls those stablecoins — Circle, Tether, or potentially a bank-issued alternative — sits at a critical chokepoint. That’s a different kind of infrastructure control than what Chalom is focused on, and it’s unclear yet how those two power centers interact.

Sharplink held 891,714 ETH as of mid-September.