Grove’s Gateway Thesis

Special thanks to Gabi, Olshansky and Arthur for their feedback


  • Hundreds of centralized RPC providers have reinvented the wheel with their infrastructure
  • Grove will be open sourcing our core gateway and quality of service logic that has served hundreds of billions of requests
  • This will accelerate expansion into new interfaces supported by POKT outside of just RPC – AI, storage, social, indexing

Shannon, POKT Network’s next major upgrade is launching later this year. Shannon will establish permissionless gateways, the lack of which until now has significantly hindered the growth of the network. Put simply, a gateway is a proxy (i.e. a smart load balancer) to the infrastructure supply incentivized by POKT Network. Anyone who needs multichain or multi-interface infrastructure at low cost can be a gateway operator. Either for your own application at scale or to serve a use case that you have deep knowledge about. The open source world is moving at an accelerating pace, and to support every new use case and capability, we need protocols like POKT.

Permissionless gateways unlock permissionless demand, and Grove will no longer need to be the leading driver for demand to the network. While the Foundation has done an incredible job onboarding new permissioned gateways, due to Morse, this process is slow, limited and nascent. In addition, the Shannon upgrade will bring about new scalability designs to lift the upper limit of the number of applications the network can handle, enabling anyone to launch their own gateway or directly connect to the network. Permissionless demand is a huge deal and has important implications for the success of the network – we will finally be removing the largest bottleneck for the growth of the demand side.

Grove has been able to emulate the expected UX of a centralized provider, but solely dependent on a web3 backend. Grove doesn’t run any nodes as an RPC provider. The Portal UI is underpinned by backend API’s, data infrastructure, quality of service, and Pocket infrastructure similar to the SDK, which interacts with thousands of nodes each running different setups across 30+ countries around the world. This is the future of all gateways – settling on permissionless networks by decoupling the stack from providing the UX/UI and quality of service from running the node infrastructure.

With the march towards open source platforms and protocols aligning incentives between hardware operators and open source services, the timing of Shannon launching couldn’t be more perfect. We are in the midst of seeing the launch of millions of rollups, L1’s, appchains, foundational open-source AI models, indexing, storage, streaming, CDN’s and social protocols like Farcaster and ActivityPub. 

In an open-source ecosystem, one company cannot handle what is effectively an infinite number of permutations for developer needs as each of these verticals grow. There are hundreds of thousands of open-source models on Hugging Face. There are hundreds of RPC providers serving niche use cases for over a thousand blockchains, and we are already seeing dozens of different LLM API providers with different value propositions for the user.

Each centralized RPC gateway has reinvented the wheel by building the same infrastructure:

  1. Load Balancing – A user and a set of permissionless backend providers
  2. Data – Protocol usage & infrastructure observability
  3. User Auth – Developers & enterprise customers
  4. Payment Rails – Off-chain payments to allow not interacting with the protocol
  5. Scaling infrastructure – Dynamically scaling to demand spikes
  6. Quality of service checks – Ongoing evaluation and ranking of provider performance

The market is saying that the demand is there, yet each RPC provider is building the same thing over and over. This is incredibly inefficient and creates friction for each provider to support new services on top of blockchain RPC. This slows down the speed at which core developers can build because providers are all focused on re-building the foundation to scale infrastructure access rather than creating value add features. Ultimately slowing down the development of the blockchain ecosystem as a whole. Grove’s is going to break this cycle.

Our thesis is to lower the barrier such that any operator can launch their own gateway on top of POKT Network. We do this by open sourcing the most difficult, and time-consuming aspects of building a gateway on POKT.

Grove’s team has spent the last 18 months optimizing what truly decentralized infrastructure looks like at scale. We have served over 700B requests atop of POKT’s permissionless network, without any direct trust, and without maintaining any backend nodes ourselves.

We will enable as many sovereign gateways as the market can support. We do this by open-sourcing our core gateway and quality of service infrastructure. Any entrepreneur will be able to spin up their own gateway that could match the quality-of-service guarantees from  Grove’s public SLA out of the box. Let any operator decide what their gateway’s value proposition is, hone in on a specific market, deploy it wherever they wish, and build a business without recreating the wheel. 

We spend 80% of our time and engineering investment on quality of service at scale. To my knowledge, there are zero DePIN participants in the market today that stick behind a public SLA that is fully dependent on a protocol that pseudorandomly chooses from a permissionless supply. We are confident enough in POKT, and our own infrastructure that Grove is willing to put our money behind the service we provide. 

We will enable more gateways as we expand into more interfaces and use-cases served by POKT Network. We will be launching AI inference with open source models hosted by POKT in the coming months – I would expect there to be hundreds, and eventually thousands of AI inference gateways serving specific niches just as blockchain RPCs expanded to indexers, oracles, websockets, relayers, rollups and other beyond L1 nodes. The same applies to every other open-source interface or protocol not yet supported by the network. 

Grove’s business is built on the same framework that we will open source. As we add new interfaces or make improvements to our core stack, everyone using the same framework will benefit from it. This is how companies in an open source environment should operate – where our core changes impact and accelerate your business built on our software. As operators adopt Grove’s open source framework, it becomes the standard by which data is read and written across the open internet – all backed by POKT Network. 

As Bill Gates said in reference to Windows, “A platform is when the economic value of everybody that uses it exceeds the value of the company that creates it.” Grove will not be successful until we see the value created by users of open source infrastructure exceed the company itself.