Kite

Kite is a Layer 1 built for autonomous software agents to transact with each other — identity, authorisation and payment designed for machines rather than for people. The premise is that agents acting on users' behalf will need to pay for services, and existing payment rails assume a human.

Existing payment systems assume a person

Card networks and bank transfers are built around human authorisation, dispute rights and identity checks. An agent making thousands of small payments for data or compute fits none of that. Blockchain payments are programmable and continuous, which matches the requirement better than anything else available.

Identity is the harder problem

An agent needs a verifiable identity, a spending mandate from its principal, and limits it cannot exceed. Without those, delegating payment authority to software is unsafe. Building identity and authorisation as primitives rather than as application logic is the substantive design decision here.

Machine payments have different economics

Agents transact frequently in small amounts, which requires per-transaction costs low enough that the fee does not exceed the payment. That constraint shapes the whole design, and it is why general-purpose chains with meaningful fees do not serve the use case.

The demand is anticipated, not demonstrated

Autonomous agents transacting at scale is a plausible near-term development and it has not yet happened in volume. Infrastructure built ahead of demand can be well positioned or early, and the distinction is only visible afterwards.

What to weigh

A coherent design for a use case with genuine logic behind it, where the market has not yet arrived and the identity and authorisation primitives are the parts worth examining closely.