Mastercard and Crossmint’s Lobster.cash have launched a new system that lets consumers authorize artificial intelligence agents to make payments using their existing Mastercard cards.
The collaboration, called Agent Pay, ties every transaction to a cryptographic “permission slip” proving that a user has allowed a specific ai agent to spend from their account. All payments still run through the Mastercard network and the security controls of the issuing bank.
How the system works
Under the new model:
- Users give explicit, cryptographically signed permission to an ai agent to make certain purchases.
- The ai agent initiates a payment, but never sees sensitive data such as the full card number.
- Mastercard’s network verifies the transaction and routes it through existing issuer controls and authentication layers.
- The card-issuing bank can apply the same fraud checks, limits, and security filters it uses for conventional card payments.
Mastercard’s chief digital officer Fourez said the partnership extends the company’s payments network into “open-agent platforms,” allowing ai-powered systems to initiate payments while staying inside established security and compliance rules.
Early access through OpenClaw
The integration will first roll out in early access via OpenClaw, an open-source platform that lets ai agents perform digital tasks, including commerce actions.
Within this framework, every completed payment is intended to be monitored by issuer-level controls, aligning automated transactions with traditional card oversight.
Lobster.cash and its blockchain infrastructure
Lobster.cash is an ai-focused payment layer built by Crossmint. It runs on infrastructure connected to:
- Solana
- Circle
- Visa
- Mastercard
- Basis Theory
- Stytch
The platform uses the Solana network to handle high volumes of low-cost transactions. According to the companies, Solana has already processed about 15 million on-chain payments from automated agents, giving the model a real-world track record rather than a purely experimental status.
This setup supports sub-cent, pay-per-use payment flows that are typically uneconomic on traditional rails, and is aimed at future machine-to-machine commerce where ai systems frequently transact in small amounts.
Crossmint raised $23.6 million in 2025, in a funding round led by Ribbit Capital, to expand enterprise-grade blockchain payment capabilities. That funding underpins the development of Lobster.cash and its role as a bridge between card networks and blockchain-based payment flows.
Addressing privacy and trust concerns
The project is designed to tackle one of the main barriers to ai in finance: data privacy and control.
Recent surveys show:
- 36% of U.S. consumers cite data privacy as their top concern regarding ai in financial planning.
- 78% of Americans use ai-powered tools in some form.
- Only 14% say they are comfortable with an ai making purchases autonomously on their behalf.
- 74% insist on reviewing any major financial decisions made by an automated system.
Agent Pay is structured to reflect these concerns:
- Ai agents operate inside strict, pre-approved spending and usage rules.
- They do not gain direct access to sensitive card credentials.
- Card issuers retain the ability to review, limit, or decline transactions as they do today.
By running payments through the same bank security layers that already govern card spending, Mastercard and Lobster.cash aim to make automated payments feel like an extension of existing protections, rather than a separate, unfamiliar system.
Growing competition around ai commerce rules
The move comes amid a broader industry race to define standards for ai-driven payments.
Visa introduced a similar approach last year with its Trusted Agent Protocol, a framework created to:
- Verify and differentiate legitimate ai agents from malicious bots.
- Support ai-based commerce within existing financial networks.
By late 2025, Visa reported hundreds of successful secure transactions initiated by ai agents through its partners using this protocol.
Mastercard’s collaboration with Lobster.cash now adds another model to this emerging segment, signaling that major payment networks are positioning themselves to handle rising machine-initiated commerce.
Why the shift is happening now
Use of ai agents in commerce is moving beyond simple chat or recommendation tools toward direct purchasing actions.
Holiday data from 2025 showed a 693% year-over-year surge in traffic from generative ai to retail sites, highlighting how quickly machine-driven activity is growing in online shopping.
Payment networks are responding by:
- Building cryptographic identity and authorization layers for ai agents.
- Reusing existing fraud, security, and compliance systems to oversee automated activity.
- Allowing developers to create new ai-driven services without bypassing traditional banking controls.
According to Mastercard’s Fourez, extending the company’s “trusted network” into open-agent platforms is meant to give developers flexibility while maintaining the level of security banks and cardholders already expect from card-based transactions.
Curious how ai transforms trading too? Explore automated strategies with our powerful ai copy trading tools and enhance your market edge today.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
