The company, founded by one of the co-founders of Circle, raised US$ 30 million to build regulated banking infrastructure for AI agents.
But the most important point isn't the valuation, nor the investor, nor the hype.
It's the thesis.
So far, much of the discussion around "payments for AI agents" has been focused on the rails: stablecoins, cards, ACH, wires, x402, wallets, and micropayments.
But a payment rail alone quickly becomes a commodity.
If the protocol is open, anyone can integrate, index, and create an interface on top.
The true moat might lie elsewhere:
In the layer of identity, governance, and control.
Catena's proposal is to give each AI agent a verifiable financial identity, linked to a responsible person or company. From there, the operator sets spending limits, allowed counterparties, authorized rails, situations that require human approval, and an auditable record of every action.
In other words: it's not enough to allow agents to make payments.
It's essential to define who can do what, on behalf of whom, with what limit, using which infrastructure, and under what regulatory responsibility.
The agentic economy won't scale just because agents can pay via APIs, reserve services, or execute transactions automatically.
It will only scale when companies, banks, regulators, and users trust that these agents can handle real money without turning autonomy into operational risk, fraud, or regulatory gray areas.
That's why the request for a national trust bank license with the OCC is so relevant.
Catena is trying to be the financial control plan for the economy of agents.
And perhaps this is one of the strongest theses for the next cycle:
Capturing permission, identity, rules, and the responsibility behind it.
Because in the world of AI agents, the most valuable asset won't just be moving money.
It will be deciding, securely, when a machine can move money on behalf of someone.
But the most important point isn't the valuation, nor the investor, nor the hype.
It's the thesis.
So far, much of the discussion around "payments for AI agents" has been focused on the rails: stablecoins, cards, ACH, wires, x402, wallets, and micropayments.
But a payment rail alone quickly becomes a commodity.
If the protocol is open, anyone can integrate, index, and create an interface on top.
The true moat might lie elsewhere:
In the layer of identity, governance, and control.
Catena's proposal is to give each AI agent a verifiable financial identity, linked to a responsible person or company. From there, the operator sets spending limits, allowed counterparties, authorized rails, situations that require human approval, and an auditable record of every action.
In other words: it's not enough to allow agents to make payments.
It's essential to define who can do what, on behalf of whom, with what limit, using which infrastructure, and under what regulatory responsibility.
The agentic economy won't scale just because agents can pay via APIs, reserve services, or execute transactions automatically.
It will only scale when companies, banks, regulators, and users trust that these agents can handle real money without turning autonomy into operational risk, fraud, or regulatory gray areas.
That's why the request for a national trust bank license with the OCC is so relevant.
Catena is trying to be the financial control plan for the economy of agents.
And perhaps this is one of the strongest theses for the next cycle:
Capturing permission, identity, rules, and the responsibility behind it.
Because in the world of AI agents, the most valuable asset won't just be moving money.
It will be deciding, securely, when a machine can move money on behalf of someone.