MoonPay is moving beyond crypto access into AI-driven trading, buying Dawn Labs and launching Dawn CLI to turn plain-language strategies into automated market execution.
MoonPay has spent years making it easier for people to buy crypto. Now it wants to shape what happens after the money lands in a wallet.
The company announced on May 11 that it has acquired Dawn Labs, an applied research startup focused on artificial intelligence and financial markets, and launched Dawn CLI, an AI trading tool built to convert plain English strategy ideas into executable trading workflows. The acquisition terms were not disclosed, which is not unusual for an early-stage AI and crypto infrastructure deal, but the direction is clear enough. MoonPay is trying to move from on-ramp provider to a larger operating layer for wallets, payments, trading and AI agents.
According to MoonPay's announcement, Dawn Labs was founded in 2025 by Neeraj Prasad, an MIT-trained computer science and engineering graduate who previously worked at Waymo, Microsoft, Citadel and Reservoir Labs. That background matters because Dawn CLI is not being positioned as a chatbot that explains markets. It is being sold as a system that researches signals, generates trading code, runs simulations and executes user-directed trades on supported venues.
That is a bigger claim than most consumer AI finance products are making. A user can describe a strategy in ordinary language, such as monitoring a prediction market based on polling movement or social data, and the tool is designed to translate that idea into code, test it and run it continuously. Dawn's own materials frame the product around prediction markets, where traders often need to react quickly to news, social signals and pricing gaps across venues.
MoonPay's core business has been the bridge between fiat and digital assets. It supports crypto on-ramps and off-ramps, payments, trading and stablecoin infrastructure, serving more than 30 million customers across 180 countries and more than 500 enterprise customers. That scale gives the company a distribution advantage if it can attach more services to the first transaction.
Dawn CLI fits into that broader stack because MoonPay has already been building tools for AI agents and developers. Its MoonPay CLI supports non-custodial wallets, token swaps, bridges, transfers, recurring purchases, limit orders, deposits, market data and fiat access. The company has also talked up MoonPay Agents, Ledger-secured hardware signing, MoonAgents Card for stablecoin spending and the Open Wallet Standard for agent wallets across chains and frameworks.
Put simply, MoonPay does not want AI agents to just hold balances. It wants them to fund accounts, access wallets, move assets, trade and spend. Dawn CLI gives that strategy a sharper consumer-facing edge because it brings AI into the trading decision and execution loop, not merely the payment flow.
The initial market is a natural one. Prediction markets such as Polymarket and Kalshi have become a testing ground for event-driven traders who watch elections, economic indicators, sports, policy decisions and geopolitical developments. The opportunity is real, but so is the friction. Many active users already stitch together data feeds, spreadsheets, scripts and exchange interfaces. Dawn is trying to compress that messy process into one command-line style product.
The risk is not just bad code
The hard question is what happens when retail users treat an AI trading assistant as financial judgment rather than software. MoonPay and Dawn can say the user directs the strategy, but the more the system handles research, code generation, backtesting and execution, the harder it becomes to draw a clean line between tool and advice.
That line will matter to regulators. Automated crypto trading has always raised questions about custody, disclosure, market access, slippage and execution quality. AI adds another layer because users may not fully understand why a model picked certain signals, how it sized positions or what assumptions were embedded in a backtest. A strategy that looks sensible in simulation can behave very differently in a thin prediction market during a fast-moving news event.
MoonPay appears aware of that problem. The company's agent infrastructure emphasizes non-custodial wallets, local key handling and user-controlled signing patterns. That helps with custody, but it does not solve every risk. A user can still authorize a strategy that trades too aggressively, relies on weak data or reacts to misleading information. In AI trading, the permission model is only one part of the safety model.
The competitive backdrop is getting crowded. Crypto exchanges and wallet companies have been adding AI assistants, trading bots and portfolio tools for years, while newer agent products promise to automate research and execution across on-chain markets. Coinbase, Gemini and other large platforms have obvious incentives to turn AI into a trading interface because it keeps users inside their ecosystems. MoonPay's angle is different because it begins with payments and wallets, then reaches toward execution.
That may prove powerful if the company can make the experience feel useful without making it reckless. The more crypto firms package AI agents as trading infrastructure, the more they will be judged on the boring details: controls, logs, simulations, execution transparency, permissions and clear user consent. Those are not side features. They are the difference between a clever interface and a compliance problem.
For MoonPay, Dawn Labs is a bet that the next phase of crypto adoption is not only about helping users buy assets. It is about owning more of the actions that follow. The market should watch whether Dawn CLI stays a specialist tool for prediction market power users or becomes the template for consumer AI agents that can move money with very little human intervention.
Also read: Microsoft's OpenAI bet now looks like venture capital at hyperscale • GitLab is cutting jobs to fund its bet on AI agents • South Korea wants citizens to share in the AI boom
