Introduction
Earlier this year, we announced Ritual Altar, a program dedicated to providing full-stack support to ambitious teams building on Ritual. Our goal with Altar was to work closely with select founders to imagine a new class of protocols that sit at the intersection of crypto and AI, leveraging our own team’s extensive expertise in both domains.
Today, we are excited to announce the first batch of Ritual Altar companies.
Welcoming New Apostles
Relic: ML-Enabled AMMs
AMMs are a core product of crypto. They enable quick capital aggregation and market formation for any asset in a permissionless manner, and have been instrumental to facilitating liquidity on-chain.
However, the reality is that most AMMs on-chain are very inefficient. Liquidity Providers (LPs) on AMMs typically suffer from losses from arbitrageurs that take advantage of “stale bids”. Furthermore, AMMs today lack expressivity around more nuanced behavior, such as screening certain types of orders, flow toxicity, and classifying participants.
Relic is building a new type of Automated Market Maker (AMM) supercharged by ML models. Liquidity pools on Relic can choose from a range of ML models accessible via Ritual’s infrastructure that can modify the underlying pool and its parameters to reach a desired goal using real-time data. From models that use CEX data to predict volatility and dynamically adjust fees, to enabling agents that can execute user orders under specific conditions, the design space here is abundant.
The Relic team has deep experience in markets, both on-chain and off-chain. They’ve worked on production scale machine learning and trading systems at places like Robinhood, Jump Crypto, Amazon and more and have actually built AMMs in other ecosystems. They’ll be using their extensive experience to tackle the challenges involved with AMMs from an entirely new angle.
Anima: Enabling Multiplayer AI
LLMs have rapidly become one of the most adopted technologies ever. They represent a breakthrough in content generation at scale, and perhaps more significantly, in enabling autonomous task automation.
It’s no secret that crypto, especially on-chain, is full of complexity. The average user needs to invest substantial time, not only in identifying which protocols to use, but how to use them. This UX experience is burdensome, and while there are considerable improvements in several domains, it still remains a large barrier to adoption.
Anima believes that LLMs and other model architectures can meaningfully improve this UX, while also enabling entirely new classes of protocols. Anima is working on crypto-native LAMs that aims to make transacting on-chain as simple as interfacing in natural language. From automating research for on-chain assets and protocols, to facilitating swapping, staking, lending and more complex strategies, Anima’s goal is to create the most performant models that help users navigate blockchains. This also includes smaller models optimized for specific use cases that LAMs can call on, such as models for yield optimization or models specific to certain consumer applications.
But Anima also believes this can be taken a step further. Traditional AI models operate in isolation, with no shared state between users. Blockchains, by design, function as shared state machines. Anima’s end goal is to merge these two and create protocols that feature multiplayer AI. Building on their work on crypto-native models, Anima is developing protocols that use blockchain primitives to enable users to collectively control models, from prompts to finetuning. This represents a leap from the “single-player” meta that is common with AI models today.
Anima’s team is deeply accomplished across both crypto and AI. Its team members have previously cofounded and raised millions for one of the largest data companies in crypto (used by over 15,000 developers and 8,000 projects), worked on AI research at top institutions (NTU, Harvard, Meta, the University of Tokyo, and more), published works at top conferences and journals (NeurIPS, ICLR and more), and, most importantly, have been on-chain power users themselves. Anima will draw on the team’s experience to develop the best AI models and protocols for on-chain users in the market.
Tithe: Universal Credit Markets
Like DEXs, lending protocols are critical for enabling robust on-chain economies. Credit layers offer users both the ability to gain exposure to different asset mixes and access yield from lending out passive holdings. To date, on-chain lending protocols have facilitated billions in liquidity through both aggregation and origination.
Yet, lending protocols on-chain are not typically as dynamic as the markets they enable. A large majority of lending protocols rely on slow and bureaucratic governance processes to control parameters like asset LTVs and interest rate models. In reality, lending protocols should be able to autonomously and quickly adapt to market conditions in real-time.
Enter Tithe, a new kind of lending protocol powered by AI models. Tithe features isolated lending pools that can integrate with models to dynamically optimize pool parameters against specific objectives. This may mean a model that constantly monitors a number of features (both on and off-chain) of the underlying assets to dynamically adjust LTV ratios – minimizing bad debt while maximizing capital efficiency. This may also mean models that can classify and assign credit scores to on-chain addresses, creating custom loan terms for individual borrowers.
By leveraging models as a core primitive, Tithe functions as a universal credit layer. Models that can accurately help price more illiquid or esoteric assets (like NFTs and RWAs) can be used to enable lending pools that extend beyond the typical ones seen. Models can also facilitate inter-pool operations, allowing pools to interact through deposits, withdrawals, and other coordinated actions. To ensure users have access to the best models, model creators on Tithe are incentivized through revenue sharing, receiving a portion of the fees generated by pools utilizing their models.
Tithe’s team features top operators and researchers that both worked on staking and credit products at a top CEX and have been extensive DeFi users themselves. Their extensive experience navigating multiple market cycles positions them uniquely to build Tithe into crypto’s universal credit layer.
Conclusion
We could not be more excited to work closely with each of these Altar teams. Their deep experiences across crypto and AI align with what we have always believed at Ritual: that the best builders are often users themselves. We will be supporting these teams across all core functionalities as they build on top of Ritual and realize an entirely new class of protocols.
If you’re interested in getting in touch with these teams, reach out to altar@ritual.net.
This is just the beginning. We’re already working with the next batch of Altar teams building novel primitives that haven’t been seen in crypto before. Stay tuned to learn more.
Interested in becoming an Apostle and building on Ritual? Apply here. As always, we look forward to seeing you at the Altar.
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.