Limited beta – contact us to request an account

About Anycast Agents

At the PlanB conference sponsored by Tether, we spoke with builders, founders, and developers who believe the future is decentralized and truly peer to peer. We share that belief. We also left with a clear view of what is missing to make that future real.

The decentralization gap we saw in the wild

Most of what people use today, from wallets to gateways to operational systems, still depends on big tech and the big cloud. Even when the asset is decentralized, the supporting infrastructure often is not.

When a single region or provider fails, the experience fails and decentralization becomes an aspiration rather than a property you can rely on.

Bitcoin is the only example of a currency that is only ruled by math that basically cannot be changed.

— Paolo Ardoino, CEO of Tether (source)

Decentralization is strongest when systems are durable, predictable, and not dependent on a small group or a single point of control.

Decentralization is also an infrastructure decision

We think decentralization is not only a protocol decision. It is also an infrastructure decision. If connectivity, event delivery, and coordination depend on a single endpoint, a single region, or a single vendor path, the system behaves centrally even if parts of it are decentralized.

  • Avoid single-region endpoints
  • Avoid single-provider dependency
  • Design for failure, not optimism

If you build technology that can only survive if you're around, that's not good technology.

— Paolo Ardoino, CEO of Tether (source)

Resilience is a prerequisite for decentralization. Systems should keep working through failures, churn, and change.

Why Layer 3 matters

Anycast is our core technology. It operates at Layer 3, in the plumbing of the internet. We work on routing, reachability, and moving information freely at global scale. That is the layer where decentralization either becomes real, or quietly collapses into a single bottleneck.

What we are building

What we believe

  • Decentralization should survive real outages, not only ideal conditions
  • Supporting infrastructure should be globally distributed, not single-region
  • Reliability and observability are not centralized control, they are operational requirements
  • The goal is fewer chokepoints, fewer hidden dependencies, and faster recovery

Publishing data, inviting debate

We will publish performance data and technical whitepapers in the coming months, including latency measurements, failover behavior, and real-world reliability outcomes. We welcome discussion and feedback from builders who share the mission of making decentralized systems work in practice.

We welcome discussion

If you are building on USDT rails or shipping encrypted agents and P2P applications, we want to talk. We are early, we are measuring everything, and we are publishing what we learn.

Interested in learning more or joining the beta?

Request Access