MoonPay Buys Dawn Labs to Bring AI Trading to Polymarket

MoonPay has acquired AI trading startup Dawn Labs and launched Dawn CLI, a product designed to let traders build and execute automated trading strategies on prediction markets using plain-English prompts rather than code. The company announced the deal on Monday and confirmed the product will launch with support for Polymarket, the largest crypto-native prediction market, with additional venues and asset classes to follow.

Plain-English Prompts in, Executable Trading Strategies Out

The product targets a clear gap in the prediction market trading stack. Building a systematic strategy currently requires a combination of development skills, quantitative analysis and portfolio management experience that most retail traders do not have. Dawn CLI is designed to compress those three roles into a single interface where users describe what they want in natural language and the system writes and executes the underlying code.

Dawn Labs founder Neeraj Prasad, now Chief Engineer of MoonPay Labs, told Decrypt that the platform addresses a specific underserved audience: traders active on prediction markets who lack the technical resources to automate their strategies. Prediction markets are one of the fastest-growing sectors in crypto, with combined monthly volume across Kalshi and Polymarket now exceeding $20 billion, and the existing tooling has not kept pace with the trading sophistication that capital flow has attracted.

The product addresses three risks that have shadowed earlier autonomous trading systems: hallucinated strategies that look reasonable but contain logic errors, unintended trades, and execution failures. Dawn CLI uses non-custodial wallets created through the Open Wallet Standard, exposes the underlying strategy code for users to review before deployment, and includes policy controls that cap how much an agent can trade, which markets it can access and how positions are sized.

Dawn Is the Latest Piece in MoonPay's AI Infrastructure Push

The acquisition is the most recent step in a broader strategy that MoonPay has been building for the better part of a year. In April, the company introduced the MoonAgents Card, a virtual debit Mastercard that allows AI agents to make purchases with stablecoins at any merchant accepting Mastercard, converting crypto to fiat at the point of sale. The company also launched an open-source wallet standard for AI agents, acquired crypto key management firm Sodot, and rolled out the MoonPay CLI as a developer interface for building agentic finance applications.

MoonPay CEO Ivan Soto-Wright framed the Dawn acquisition as fitting into the company's four-pillar strategy of "fund, tokenize, trade and spend," with Dawn slotting into the trade layer. Prasad described the broader thesis bluntly: MoonPay does not view AI agents and human users as separate customer bases. The infrastructure is designed for both, with a human setting the strategy and the agent executing it.

The move sits within a wider industry trend. Gemini launched an agentic trading feature for AI agents in April. Coinbase, Stripe and Amazon have all launched AI-focused stablecoin payment rails in recent months. Solana and Google announced similar agentic stablecoin infrastructure. The shared assumption across these initiatives is that autonomous software will become a significant share of transaction volume in crypto markets, and the infrastructure to support that needs to be built now rather than retrofitted later.

Is Polymarket the Obvious Place to Start?

The choice of Polymarket as the launch venue makes commercial sense. The platform recorded $10.57 billion in trading volume in March 2026 and counted nearly 478,000 monthly active wallets at its peak last October. It operates on the Polygon blockchain with USDC as the settlement currency, which makes it natively compatible with the kind of non-custodial wallet architecture Dawn CLI uses. MoonPay already has an established commercial relationship with Polymarket as its fiat onramp partner, which provides a distribution channel for the new product.

Whether plain-English prompts can reliably translate into well-executed prediction market strategies will depend on factors the product launch does not resolve. Strategy quality is constrained by the user's understanding of what they actually want to do, not just their ability to describe it. The policy controls Prasad described will protect against catastrophic losses but will not prevent users from deploying flawed strategies that lose money in less dramatic ways. The more interesting test is whether Dawn CLI can attract traders who would otherwise have stayed manual, rather than simply giving existing algorithmic traders an easier interface.