The Cleric Speaks - Unveiling the Ritual Chain
Today, we are thrilled to share the first comprehensive look at the Ritual Chain architecture.
The team behind Ritual has been hard at work for two years, researching, designing, building, and testing a novel layer 1 architecture with a simple goal: enhance what developers can do onchain today to attract the users of tomorrow.
In this post, we will discuss the current state of the AI industry, how its structural issues gave rise to the nascent Crypto x AI landscape, and explore the shortcomings of existing layer 1s in servicing this intersection. We’ll also provide a high-level overview of the Ritual Chain’s design, explore what’s possible on Ritual that isn’t possible elsewhere, and map out our vision for how Ritual enables the future of expressive onchain computation for AI and beyond.
This is the first in a series of posts exploring these topics (and others) in more depth, follow along at RitualFND on X or join our Discord to stay up to date.
Scheming Viziers - Gatekeeping and Obfuscation of AI
When Ritual started in 2023, the team was certain about 2 things:
(1) Crypto and AI are the two most important and fastest growing technologies of our time.
(2) They are structurally very different (decentralized vs. centralized, open vs. closed source, etc.) but have immense synergies.
Unfortunately, the current status quo of AI is not sustainable, equitable or desirable. Closed-source models are trained on secretly scraped or purchased data. Outputs are produced opaquely, without any guarantees of verifiability. Datasets and compute are in such high demand that the markets serving them have become monopolistic and exclusive. This paradigm both stifles innovation and centralizes control over a generational technology in the hands of a select few. AI is an inevitability, but on its current course, the future looks dim.
However, the properties enshrined by blockchain technologies offer the possibility of an alternate future in which the next generation of expressive, heterogeneous software can be developed, deployed, and natively run in decentralized, transparent, and verifiable environments. With the help of sovereign blockchain infrastructure, the path to AGI can be democratized, decentralized, and accessible to all.
It’s clear that these synergies, and the problems they address, are broadly felt: over 1000 projects building at the intersection of AI & blockchains have sprung up in the past few years, generating enormous value. At the time of this writing, the total market capitalization of agents and agent frameworks alone is ~$6.2B, and the broader Crypto x AI category is estimated to have a market capitalization over $20B.
Yet, current Crypto x AI solutions are lacking. Most of the infrastructure developed still has centralized chokepoints, reintroducing censorship and single point of failure risks. They also tend to suffer from a lack of composability across the stack, such that smart contract developers are unable to access the same model from different applications, not to mention from different chains. The vast majority of agents are not doing anything useful onchain, or even offchain (earning them the Crypto Twitter moniker: “slop reply bots”), and no one even believes they’re actually agents and not humans behind the scenes. Developers that want to use AI models for onchain use cases have to jerry-rig their own stack, a process which should ideally be abstracted away and handled at the protocol layer.
Enter Ritual: a layer 1 blockchain purpose-built for AI and beyond. With Ritual, users can actually bring AI models and other forms of expressive compute onchain.
A Candle in the Crypt - One Final Blockchain?
Ritual's early products like Infernet (the first oracle network for AI) and Frenrug (an experiment showcasing a multiagent setup taking actions onchain ) were successful proofs of concept, but also quickly revealed that existing blockchains were not practical deployment environments for AI. By nature, many layer 1 blockchains are architected to enable the same computation to run on (mostly) the same set of hardware across a distributed network. But infrastructure in AI doesn't follow this path—different models and architectures require different types of hardware depending on the workload. It would be overkill to run a regression model on an H100, just as it wouldn't make sense to run a large parameter LLM on a CPU.
The current crop of layer 1 blockchains was just not purpose-built to enable any sort of expressive computation beyond the capabilities of existing smart contract platforms. In practice, there are certain types of computation, like AI model operations, that simply aren’t feasible to run in replicated settings like those typical to blockchain architectures. A chain that provides native access to a variety of these types of compute (which we call expressive compute) significantly expands the design space for onchain use cases.
These revelations sparked the research and development of what would ultimately become the Ritual Chain: a new layer 1 meant to enable expressive onchain compute, allowing users to access net-new compute primitives like AI in a single, unified environment. In turn, this enables a set of applications only possible on Ritual.
Let’s dive into how it works.
Rites of Abiogenesis - The Ritual L1 and a New Path Forward
Ultimately, the team realized the primitives composing this new architecture presented new avenues for both builders and end users alike. A chain purpose-built for AI is also the world’s most expressive blockchain, paving the way for an entirely new ecosystem of sophisticated onchain applications. Some of these applications will undoubtedly be built with, or powered by, AI models. Some of them will not. All of them will run seamlessly and verifiably on Ritual.
The alchemy of our experiences, insights, and values produced the novel architecture of the Ritual Chain, whose full design can be found in our Developer Docs. Some of the more novel features you’ll find on the Ritual Chain include:
○ Node Specialization: Nodes can selectively choose, and be appropriately rewarded for, the type of workloads they want to service based on their unique computational and proof generation capabilities. The network therefore incentivizes both high-performance and resource-constrained users to participate, and allows for granular optimization resulting in increased efficiency and flexibility for users.
○ EVM++: A backwards-compatible extension of the Ethereum Virtual Machine (EVM) with expressive compute precompiles that enable access to EVM++ sidecars, native scheduling and enshrined oracles, and built-in account abstraction.
○ EVM++ Sidecars: Sidecars are dedicated execution environments that allow Ritual to expose expressive computers like AI while keeping the onchain state lightweight. For example, sidecars are used to run model inference outside of the execution client, allowing it to run and scale in parallel to the core execution layer. Currently supported sidecars include Classical ML & LLM Inference, ZK Proving & Verification, TEE Code Execution, and Chain Abstraction.
○ Modular Computational Integrity: Ritual flexibly supports a number of methods to prove computational integrity of model execution. Zero-knowledge machine learning, TEE attestations, optimistic machine learning, and probabilistic proof machine learning are all supported natively and agnostically.
○ Verifiable Provenance: Enshrines the ability for model creators on Ritual to prove ownership of their models in the wild, expanding distribution pathways and unlocking new structures of incentivization and fee sharing through a set of new primitives, starting with vTune.
○ Resonance: A state-of-the-art fee mechanism unique to Ritual that allows for optimal pricing of transactions across heterogeneous hardware by introducing a two-sided marketplace between compute providers and users. This lets Ritual price the types of compute offered on the chain (like AI models) under one market.
○ Symphony: Symphony accelerates consensus mechanisms by building upon an Execute-Once-Verify-Many-Times (EOVMT) model in which select compute nodes act as sole executors of workloads, generating succinct sub-proofs for executed compute. This allows the network to reach consensus on resource-intensive computational outputs (like model inferences) without needing to replicate execution across every node.
○ Scheduled Transactions: Allows for enshrined transactions, where a user can send a transaction once and it non-interactively gets called every block, based on a set of conditions that users specify. Aside from being able to migrate all offchain oracle setups today onchain, this enables agents to live onchain where they can run forever without anyone being able to take them down, which isn’t possible on any other chain today.
Ritual unites all of these features and more under one roof.
So what can you actually build on Ritual?
With AI natively enshrined onchain, protocol builders can harness models as core features, rather than just tokenized wrappers. This enables next-generation DeFi primitives like AI-pegged stablecoins, LLM-driven prediction markets, basis trading, and dynamic lending/borrowing. Agents can now take onchain actions seamlessly to interact with protocols, enabling both improved UX for users via natural language interfaces and the formation of new types of onchain consumer applications based around these agents (we’ll be diving deeper into agents on Ritual in a future post, so keep an eye out!). We’ve shared builds of some of our ideas in our Tutorials, and are already working with a number of talented ecosystem teams to build out applications at the frontier of Crypto x AI, but we’re most excited to see what killer use cases our community of builders comes up with. If that’s you, reach out to us via Twitter or message us on Discord.
Ritual’s primitives also enable net-new products beyond Crypto x AI. These include:
○ Hypercustomizable RaaS: Rollups that use EVM++ and Resonance can easily implement and price specialized precompiles that give their users access to expressive compute without overloading the sequencer. This could mean specialized, enshrined oracles for new markets (i.e. RWAs), TEE & ZK sidecars for sensitive operations like asset settlement, or a set of AI models accessible by all users on the rollup.
○ Generalized Prover Networks: Existing prover networks typically lack the ability to price special-purpose proof systems, leading to centralization in prover markets. Ritual enables an integrated prover marketplace that can price different proving systems across heterogenous hardware using Resonance.
○ Model Marketplaces: With primitives like vTune, users can prove attributions to both model creators and data providers, allowing marketplaces for models with verifiable onchain monetization and the ability to buy, sell, and borrow models. Graph operations over related model IP can be run and priced efficiently using EVM++ and Resonance, while integrations with Arweave/HuggingFace provide censorship-resistant storage options.
Each component of the Ritual stack deserves its own exploration of its capabilities and how they enable these applications and more. We’ll provide more detailed breakdowns in future posts.
Oracles of Light - Summoning the Future
While other chains continue to focus on improving the scale and speed of what’s currently possible, we envision a chain that will enable what’s currently _impossible.
It will be a chain that gives users access to a variety of computational resources, efficient and specialized systems of resource allocation, and diverse verifiability and parallelization mechanics.
It will provide enhanced customizability and composability in a familiar development framework, offering flexibility and native integration to developers and end users alike.
It will symbiotically benefit from seamless connectivity with other ecosystems, providing resources and leveraging existing applications.
It will be built by a deep network of community members, builders, partners, and collaborators—a collective undertaking to create an ecosystem focused on bringing net-new applications to crypto as a whole.
It will be Ritual.
Disclaimer: This post is for general information purposes only. It does not constitute investment advice or a recommendation, offer or solicitation to buy or sell any investment and should not be used in the evaluation of the merits of making any investment decision. It should not be relied upon for accounting, legal or tax advice or investment recommendations. The information in this post should not be construed as a promise or guarantee in connection with the release or development of any future products, services or digital assets. This post reflects the current opinions of the authors and is not made on behalf of Ritual or its affiliates and does not necessarily reflect the opinions of Ritual, its affiliates or individuals associated with Ritual. All information in this post is provided without any representation or warranty of any kind. The opinions reflected herein are subject to change without being updated.