eToro Agentic Trading Users Jump 250% as Crypto Trades Fall 62%

eToro disclosed on Oct. 8 that the number of users engaging with API and agentic trading on its platform rose 250% quarter over quarter, the same day the company announced connectors for Claude, ChatGPT and Grok that let customers link their own artificial-intelligence agents to investment accounts. The surge in machine trading arrived alongside a sharp contraction in the crypto activity eToro identifies as the main driver of its earnings: October crypto trades fell 62% year over year to 2 million, with the average amount invested per trade down 29% to $228.

The company's third-quarter figures, for the period ended Sept. 30, point the same way. Net contribution came in at $186 million, 13% below the level of a year earlier, and eToro attributed the shortfall mainly to lower crypto trading. The 250% surge in agentic users therefore sits on top of a shrinking trading base rather than a growing one.

What did eToro disclose?

The announcement carried five numbers worth holding on to:

• Users engaging with API and agentic trading: up 250% quarter over quarter.
• Connectors for Claude, ChatGPT and Grok: announced Oct. 8, after the third quarter closed Sept. 30.
• October crypto trades: 2 million, down 62% year over year.
• Average amount invested per crypto trade: $228, down 29% year over year.
• Third-quarter net contribution: $186 million, down 13% year over year, which eToro ties mainly to lower crypto trading.

A 250% quarterly increase means the cohort of agentic users now runs roughly three and a half times its prior-quarter level. The company published no absolute count for that cohort, so the segment's size cannot be checked independently, and the disclosure does not separate API-driven activity from agentic activity. In plain terms: adoption of machine trading clients is accelerating, and the size of the group doing it remains unstated.

What do the connectors do?

The connectors let users connect their own AI agents to the platform's investment accounts rather than relying on eToro's own interface alone. A second feature, Agent Portfolios, extends that arrangement to a subaccount inside a user's portfolio, which lets a model be pointed at a defined slice of holdings instead of the whole account. The operational consequence is that order routing and allocation decisions can now be handed to third-party models, including Claude, ChatGPT and Grok, whose behavior sits outside eToro's own product design. The announcement did not specify which models customers use most, whether execution is constrained by any rule set, or what disclosure and suitability standards apply to agent-directed orders.

How much crypto trading is there to grow?

Across a 31-day October, 2 million crypto trades amount to roughly 65,000 a day, and the 62% year-over-year decline implies an October of about 5.3 million trades a year earlier. Multiplying the $228 average ticket by the monthly trade count yields roughly $456 million of capital deployed across those trades, a figure eToro has not itself published.

Two timing caveats matter more than either calculation. The connectors went live on Oct. 8, so October contains only about three weeks of connector availability; the 62% decline cannot be read as a verdict on the new tools. And the trade data covers October, the first month of the fourth quarter, while the 13% contribution decline describes July through September. The crypto slowdown predates the connectors, and October extends that slowdown into the new quarter. The company does not disclose crypto's share of net contribution, so no one can size precisely how much of the $186 million depends on the asset class the agentic tools target.

What does the 250% figure leave unanswered?

Three gaps deserve scrutiny. First, the base: without an absolute user count or a comparison against eToro's total customer base, a 250% gain cannot be weighed against the platform's scale. Second, monetization: the disclosure reports user counts and trade metrics, but not how much revenue, deposits or contribution the agentic cohort produced. Third, causality: nothing in the announcement shows that users adopted agentic trading because of the connectors rather than in anticipation of them, or that those users trade crypto at all.

The restrained reading is that eToro is selling the plumbing while the retail crypto cycle works against the volume that plumbing was built to capture.

What to watch next

The first clean test arrives with fourth-quarter and full-year results, the first window in which the connectors will have been live for a full quarter rather than three weeks. The metrics to watch are the two that are already moving in opposite directions: the count of agentic users, and the average invested per crypto trade, which at $228 was down 29% year over year.

The next reporting deadline, the fourth-quarter and full-year release, will be the first point at which agentic adoption can be measured against funded trading rather than against user counts alone.