Bitget AI Hits 1 Million Users and $1.2B in Agent-Routed Trades — Exchanges Just Found Their Next User-Acquisition Loop

The phrase that buried the headline in Bitget's May 15 press release was four words long: "from chat to execution." That is what CEO Gracy Chen called the shift the exchange wants to claim, and it is the cleanest articulation any centralized venue has offered for what an "AI-native exchange" actually means in 2026. The press release itself read like a product roll-up — 58 AI tools, 1 million users, $1.2 billion in agent-routed volume — but the operational story underneath is more interesting. Bitget has found something that the other top-five exchanges have not yet matched: a user-acquisition loop where the AI feature is the funnel, not the upsell.

From signals to execution: what actually shipped

For most of 2024 and 2025, exchange "AI features" meant two things: a chat box that explained market events, and a portfolio assistant that summarized P&L. Both were marketing wraps on a hosted language model — useful for retention surveys, marginal for revenue. Bitget's May 15 unified launch reframes the category by surfacing four operational products that all touch order flow directly.

GetClaw is the entry point: a zero-install conversational agent that scans Bitget's order book and external data, summarizes movements, and proposes trade ideas. The user does not have to download anything; the agent lives inside the exchange UI. GetAgent is the execution layer — natural-language strategy ("buy ETH if it breaks $4,200 with 3x leverage and a 5% trailing stop") converts directly into live orders and managed positions, with the agent handling rebalancing and stop-loss management in the background. Agent Hub is a developer surface with open APIs that lets third-party agents — including hosted LLMs or custom code — read Bitget data and submit orders, intended as the long-tail expansion. AI Trading Playbooks in beta lets a user describe a strategy in plain language, backtest it on Bitget's historical data, and then deploy it as a hosted bot, with an emerging marketplace for top-performing playbooks.

The press materials reference "58 AI-powered tools," but Bitget never enumerates all 58 publicly. From the named components above plus the company's earlier product blogs, the 58 figure most likely covers analytical sub-features (chart-pattern detector, news summarizer, on-chain signal feeds) bundled inside GetClaw and GetAgent — closer to "feature count" than discrete products. That detail matters because it suggests the headline number is engineered for narrative, and the substance lives in the four named pillars.

Why exchanges suddenly care about agent flow

The competitive pressure on Bitget is well-defined. Binance still dominates spot and derivatives volume but has not given its grid-bot and arbitrage suite an "agent" brand. OKX shipped Agent Trade Kit — an open-source natural-language strategy framework that integrates with its Onchain OS — but kept the framing developer-first. Bybit's Aurora AI is more sophisticated in strategy classification but lacks the conversational front door. Bitget's bet is that a single coherent brand ("the AI trading exchange") is worth more than a sprawl of features that no retail user will ever string together.

The funnel math is where it gets interesting. A user who installs an agent that runs strategies overnight is not the same user as one who manually trades in the morning. The agent generates fee volume around the clock, on chains the user has never thought about, and produces lock-in not through loyalty but through configured state — backtested strategies, saved signals, rebalancing rules. That makes user acquisition compound in a way classic exchange features did not. Bitget's own May 14 UEX retail report — separate from this announcement — noted that 51% of its users already touch an AI feature. If that adoption curve continues, the AI suite is no longer a side product. It is the funnel.

The exchange comparison is also a market-structure story. Bitget closed 2025 ranked #6 globally by spot share with ~6.4% of volume and ~45.5% year-over-year growth, putting it neck-and-neck with Bybit and gaining on OKX. The $4B/day derivatives book is roughly $1.4 trillion annualized; $1.2B cumulative on AI tools is, by that measure, well under a day's flow. But Bitget is pointing at a future where that ratio inverts as agents become a primary trading interface for retail.

The honesty test for agent-routed exchanges

The unflattering reading of Bitget's numbers is that "$1.2B cumulative" is exactly the kind of figure designed not to be benchmarked. No start date, no breakdown by product, no separation of natural retail volume that simply happened to flow through an AI button versus genuinely autonomous agent decisions. Anyone who has worked on crypto-exchange product knows that "powered by AI" can mean almost anything internally, from a fully autonomous strategy bot to a rules engine with a language-model wrapper. The story Bitget is telling — that AI is now driving execution — is plausible. But the company has not yet published the kind of breakdown (per-product MAU, average position duration, agent-only PnL) that would let an outsider verify it.

Two specific risks are worth flagging. First, regulatory exposure. Several jurisdictions are still working through whether AI-driven trading on retail-facing exchanges counts as managed-account activity, with the licensing implications that come with that. Bitget's official channels describe GetAgent as user-configured, but the line between a configured-by-user strategy and a managed product gets thinner the more autonomy the system has. Second is the failure mode. Hosted strategies turning over user funds at machine speed will, eventually, run into a market dislocation that wipes accounts faster than support tickets can be opened. How an exchange handles that — and whether the agent surface accelerates contagion or contains it — is the next operational headline to watch.

Beyond the headlines: Bitget is not alone in this race — it is simply the first to package the story coherently. Expect OKX, Bybit, and eventually Binance to consolidate their own AI surfaces under a single brand within the next two quarters. The interesting test will not be whose AI is most capable. It will be whose data-and-execution loop attracts the most third-party developer agents. The exchange that wins Agent Hub gets the next decade of fee volume that no human is initiating — and that, more than the 1-million-user banner, is the actual prize Bitget is going after.

Reviewed by Jason Lee, Founder & Editor-in-Chief, BlockAI News.

Sources

Primary sources

• GlobeNewswire — Bitget Introduces Unified AI Trading Ecosystem (May 15, 2026)
• Bitget — 1M users / $1.2B official news
• Bitget — GetAgent AI Trading Assistant launch
• Bitget blog — GetAgent major upgrade
• CoinGecko — Bitget futures volume reference
• @BitgetGlobal — Bitget official X account
• @GracyBitget — Gracy Chen, CEO

How we report: This article cites primary sources, regulatory filings, and on-chain data where available. BlockAI News uses AI tools to assist with research and first-draft generation; every article is reviewed and edited by a human editor before publication. Read our full How We Report page, Editorial Policy, AI Use Policy, and Corrections Policy.