Skip to content
Advertisement
Coinbase CEO says AI agents could handle nearly all transactions within a decade

Coinbase CEO says AI agents could handle nearly all transactions within a decade

Coinbase official logo (public domain, Wikimedia Commons) — CryptoBriefing brand treatment

Brian Armstrong pitched 'agentic finance' in Austin, reiterated a $400,000 Bitcoin target for 2030, and revealed how much prediction markets now matter to Coinbase

Brian Armstrong thinks the next big customer base for crypto won’t have a pulse. Speaking with CoinDesk at the TomorrowX conference in Austin on Tuesday, the Coinbase CEO said AI agents could end up handling the overwhelming majority of all transactions within a decade.

His reasoning is simple. AI agents will soon outnumber humans, he argued, and they will need some way to pay each other for goods and services.

Agentic finance gets a name and a price tag

Armstrong has a label for this vision: agentic finance, or AiFi. The idea is that autonomous AI agents will carry out financial transactions on behalf of people, and increasingly with one another.

Armstrong attached a big number to the pitch. He projected that AI agents could facilitate between $3 and $5 trillion in transaction volume by 2030.

He also made clear which rails he thinks those agents should use. In his view, crypto, particularly stablecoins and Coinbase’s Base network, is the best infrastructure for fast, flexible payments that work across borders.

Advertisement

Bitcoin, prediction markets and tokenized stocks

Armstrong didn’t stop at AI. He said he believes Bitcoin has hit its bottom for the current cycle and could reach $400,000 by 2030, a target he has put forward before.

The more concrete numbers came from inside Coinbase itself. Prediction markets and derivatives now make up 25% of the exchange’s trading fee revenue, which the research places in the third quarter.

Prediction markets alone reached an annualized run-rate of $200 million after launching earlier in the year.

Tokenized stocks are also picking up speed. Armstrong said shares traded on Base have surpassed $5 billion in volume. He went a step further and urged public companies to seriously consider tokenizing their own equity.

Why Coinbase is leaning into this story

Coinbase has spent years trying to be more than a place where people buy and sell coins. Trading fees tied to volatile spot markets make for lumpy revenue, swinging hard with crypto sentiment.

The numbers Armstrong shared show the diversification is working, at least on paper. A quarter of trading fee revenue now coming from prediction markets and derivatives means Coinbase is less dependent on whether retail traders feel like buying altcoins this month.

What this means

For investors, the clearest takeaway is the revenue mix. Prediction markets hitting a $200 million annualized run-rate and derivatives contributing meaningfully to fees suggest Coinbase is building income streams that may hold up better when spot trading slows.

The AI agent thesis is longer-dated and harder to verify. A $3 to $5 trillion projection for 2030 is Armstrong’s estimate, and it depends on several things going right: AI agents becoming trustworthy enough to handle money, regulators getting comfortable with machine-to-machine payments, and businesses choosing crypto rails over upgraded traditional systems.

The tokenized stock push carries its own stakes. Crossing $5 billion in volume on Base shows there is appetite, but persuading public companies to tokenize their shares would require legal, custody and exchange questions to be resolved.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.