Skip to main content

Networks are fragile.

Pocket is unstoppable.

High performance self-healing access to any public data anywhere.

Hundreds of millions of requests served today.

Data is exploding and so are our infrastructure needs.

  • 2024 saw 2.5 million exabytes of traffic monthly transferred. The equivalent of over 5 trillion hours of video.
  • AI, autonomous agents, and novel protocols are pushing current infrastructures to their limits, exposing bottlenecks and single points of failure.
  • A public, open-source foundation is needed—one that ensures data flows freely and reliably across applications, businesses, and communities.

Trusted by Leading Builders, Innovators, and Communities

  • Over 10,000 nodes powering data access for global AI and Web3 projects
  • Almost 1 Trillion relays served to date across diverse protocols
  • Hundreds of developers actively contributing, building and using Pocket Network

 

Gateways

Spin up custom gateways to provide seamless access to diverse data streams, including specialized AI models and emerging blockchains.

Applications

Access the limitless APIs of the open internet to power your application.

Nodes

Run the permissionless and censorship-resistant infrastructure for the open internet.


Buy POKT

POKT is the lifeblood of the network, coordinating global infrastructure. Purchase POKT to stake and participate in the network.

Latest News

Wake Up Call: Decentralization is not a meme

As everyone knows by now, AWS fell over for about a day this month, resulting in what’s likely to be billions of dollars in losses. Countless high availability applications and platforms blew out their annual SLA requirements in a single downtime incident. You’re only allowed 52.6 minutes of downtime per year to achieve four nines,…

Blockchain Governance Models Explained

Most people don’t buy a token because of governance. They buy the story, the product, and the community. That’s fine. But once you’re holding the bag, the stuff that keeps value from dumping, maintains pricing, supply discipline, and reliable access comes from how a network is governed. In other words, you don’t need governance to…

Open-Weight AI Meets Open Access—Auditable Inference with Permissionless API Gateways

In August 2025, OpenAI introduced two downloadable open-weight models, gpt-oss-120b and gpt-oss-20b, that you can run on your hardware or inside your cloud. The 120b variant is engineered to run on a single 80 GB GPU (e.g., H100/MI300X) thanks to native MXFP4 quantization. In contrast, the 20b is tuned to fit in roughly 16 GB…