Six groups of companies are currently competing to control how autonomous AI agents pay for things, verify identity, and execute transactions. This race will determine who owns the infrastructure layer for agentic commerce. The rental economy has no seat at that table, and the consequences of that absence will be felt for a generation.
The contest breaks along roughly six camps: OpenAI and Stripe are building an agent-to-commerce protocol called ACP. Google and Shopify are developing a Universal Commerce Protocol. Google, Mastercard, and PayPal are collaborating on AP2, which introduces the concept of agent-to-agent payment mandates. Coinbase and Stripe are promoting x402, a micropayment protocol using stablecoins over standard HTTP. AWS and the cloud providers are building agent governance runtime environments. And a sixth group is focused on liability and evidence standards for agentic transactions.
Each camp is designing infrastructure for roughly the same use case: an AI agent that can discover a product or service, verify purchasing authority, execute a transaction, and complete the loop with a confirmation. Buy, pay, confirm. That is the model they are optimizing for.
The rental transaction is not a buy button. It is a multi-year financial relationship with mandatory identity verification, insurance requirements, credit reporting obligations, and ongoing payment obligations. None of the six camps are building for this. The protocols being designed for e-commerce do not contain the governance structures that rental requires.
What makes rental different
A hotel booking can be handled by any of the six emerging protocols. A grocery order can be handled by any of them. A rental lease cannot, and the reasons are specific.
Rental involves a mandatory identity check with legal standing. In most jurisdictions, landlords are legally required to screen tenants in ways that are not required for other commercial transactions. The screening decision is governed by fair housing law, privacy law, and consumer protection law simultaneously. An AI agent executing a lease is making decisions that carry legal weight that a coffee order does not.
Rental involves embedded insurance. In an increasing number of jurisdictions, renter's insurance is mandatory as a condition of tenancy. That insurance requirement needs to be verified, placed, and maintained as part of the rental transaction, not as a separate workflow. None of the current agentic commerce protocols have a mechanism for embedded financial products within a transaction.
Rental involves ongoing financial obligations, not point-in-time purchases. Lease terms run twelve months or more. Payment obligations recur monthly. The relationship between tenant and property management company continues for the duration, with obligations on both sides. Dispute resolution may be required at any point during that relationship. Buy-pay-confirm protocols are not designed for any of this.
Rental households in the US alone. More than one-third of all housing units are occupied by renters.
Competing protocol camps racing to control the agentic commerce infrastructure layer, none of them designed for the rental use case.
Annual rent collected in the US. One of the largest recurring financial flows in the consumer economy, and largely invisible to the agentic commerce discussions.
Why this matters now
Agentic commerce infrastructure gets built once, or more accurately, the first version of it becomes extremely difficult to displace. The protocols that win this race will govern how AI agents transact for the next decade. The rental economy is not in this discussion at the protocol design level.
The risk is not that AI agents will be unable to handle rental transactions. They will handle them, using whatever protocol exists. The risk is that they will handle them using infrastructure that was not designed for the specific requirements of tenancy: regulated identity, embedded insurance, bilateral evidence, and multi-year financial relationships.
When that happens, the deficiencies will be discovered through failures: lease disputes without adequate evidence records, insurance gaps not caught at the point of transaction, identity verification that satisfies a commerce protocol but not a fair housing obligation. These are foreseeable problems. The time to address them is before the protocols are set, not after the first wave of failures surfaces.
The rental economy needs a protocol layer designed for its specific requirements. That means engaging with the infrastructure conversation that is happening now, not after the general-purpose protocols have already been deployed at scale and hardened into standards.