Cover image generated via Hyperbolic

Liquid Token Review — OLAS

TL;DR

• Olas is showing promising signs of growth, largely driven by its prediction markets agent, Olas Predict.
• We are keeping an eye on OLAS’s economics because of the high inflation and relatively low inflows (income and user demand) in its current state. OLAS is a low-float, high-FDV token.
• If AI agents explode and major new use agent cases emerge on Olas, this will ignite the growth flywheel and propel OLAS to new heights. Olas also stands to benefit as an index play for the AI agent narrative. Meanwhile, investors would be wise to stay sharp and watch for signs of momentum.

Source: Olas

AI agents feel a lot like the internet in the ’90s—there are skeptics everywhere, but soon, everyone and every company will have their own agent.

Soon enough, we will live in a world where AI agents outnumber humans—a hundred billion of them.

It’s also quite likely that AI agents will be handling the majority of transactions on the blockchain instead of humans.

AI agents are strong at processing unstructured data and adapting to dynamic environments with minimal human input. They’re getting smarter by the day, steadily improving in reasoning and planning tasks.

Olas (rebranded from Autonolas) is positioning itself for this agentic future.

Their bold mission: to unleash an ocean of autonomous AI agents, all owned and powered by users like you and me.

Olas has been around for a while. Plenty has been written about them, but there’s one thing missing—a fresh, up-to-date report covering the latest developments through Q4 2024. So, we rolled up our sleeves and did the research ourselves.

In this piece, we’ll explore what OLAS has been building, its current traction, and how it impacts the OLAS token.

A Brief History and Current State

Founded in 2022, Olas has been consistently rolling out new features. 2023 was a pivotal year, marked by the release of their whitepaper in February, their TGE in July, and surpassing 100,000 transactions for their AI agents by September 2023.

Fast forward to 2024, and Olas is now deployed on eight blockchains: Ethereum, Polygon, Gnosis, Solana, Arbitrum, Celo, Optimism, and Base. Notable milestones include crossing $1 million in protocol fees by January 20th and, by June 8th, accounting for 20% of all SAFE transactions on the Gnosis chain.

Source: Dune

The chart above illustrates the steady growth of autonomous service transactions on Gnosis since May 2023, showing a consistent upward trend. By September 2024, weekly transactions hit record highs, with >1.2M cumulative transactions since the start.

But let’s dig deeper to understand what’s driving these numbers and what they reveal about OLAS’s future — should we be bullish or bearish?

#1: The Olas Product Suite

In this section, we’ll provide a high-level overview of the product you can grasp in just a few minutes. If you’d like to dive deeper, you can check out their whitepaper

To achieve its vision of “an ocean of AI agents”, Olas provides a robust suite of AI agent products and services, building a framework necessary for large-scale deployment.

One of the challenges with bringing AI agents on-chain is that they must run continuously, which can be computationally expensive.

Smart contracts, while powerful, aren’t practical for this purpose—they’re constrained by processing capabilities (memory, cost) and can’t access data outside their native blockchain (like external APIs).

Olas addresses this by using an off-chain system to handle the agent’s logic, paired with an on-chain presence, typically a wallet.

This synergy between off-chain and on-chain systems is the heart of the Olas platform…

…The Open Autonomy Framework.

Olas has developed a software framework for creating and managing decentralized autonomous agents. As far as we know, this is one of the few open-source agentic frameworks that is explicitly crypto-friendly.

Source: Autonolas

The framework includes a Multi-Agent Systems (MAS) Architecture that enables multiple agents to operate autonomously while working together to achieve a common goal. Each agent handles different tasks, making decisions based on its environment or data shared with other agents. This decentralized decision-making ensures the system can function without a central authority.

The MAS framework is modular and flexible, allowing developers to customize agents by combining components like skills, protocols, and connections to fit specific services. While agents themselves operate off-chain, their decisions and actions are anchored on-chain for transparency.

Olas agents work in conjunction with a consensus node, and each agent runs an application within a finite state machine (FSM) replicated on a temporary blockchain. Agents achieve consensus on this temporary blockchain, allowing them to coordinate actions.

Key building blocks include:

• Consensus Gadgets: These enable multiple agents to agree on decisions or states without logging every detail on-chain. Olas currently uses Tendermint as its consensus engine but can also support alternatives like Avalanche or Solana. For instance, agents might use consensus gadgets to agree on a token’s optimal price, only recording the final outcome on-chain.
• Finite state machines (FSMs): FSMs govern agent behaviour, with each agent operating through predefined states. For example, an agent monitoring token prices might cycle through states like “fetch price,” “analyze price,” and “submit result,” transitioning between these states based on specific conditions. \

Source: Autonolas

• Autonolas Protocol: This is the system through which agent activity is anchored on the blockchain. This involves several smart contracts and NFT implementations, ensuring that the actions and decisions of Olas agents are securely recorded on the target settlement layer.

Source: Autonolas

By combining these components, Olas has enabled the development of various AI agents that can interact both off-chain and on-chain. The result is a robust framework that supports a decentralized AI agent economy.

What Agents Are People Building?

Here are the AI agents we’ve found that are created on Olas:

• AI
  • Mechs - performs off-chain tasks like LLM requests and returns the results on-chain. Some have likened it to an on-demand brain for applications.
  • Prediction Agents - AI agent designed to trade in prediction markets
• Oracles
  • Price Oracle - Fetches BTC prices from various sources to provide accurate pricing data.
  • ML APY Prediction Oracle - Uses machine learning to predict the future APY of Uniswap v2 liquidity pools.
• Messaging
  • Olas messaging - A communication network that allows agents to find each other and exchange messages using wallet addresses
• DeFI
  • Autonomous Fund - A Balancer smart-managed pool that follows an autonomous strategy based on the crypto fear and greed index. It buys when fear is high and sells when greed is high.
• NFTs
  • El Collectooorr - Bids on new Art Blocks NFT mints and manages them in vault
• DAO Tooling
  • Olas Contribute - An autonomous service that tracks community interactions. Agents monitor activity, record contributions, reward members who contribute the most, and maintain a leaderboard

You can see that most agent transactions thus far are from Trader and Mech agents, largely tied to the prediction markets service.

Pearl, A Better UX for Running Agents

In June 2024, Olas released Pearl, a simple and intuitive interface for users to get involved with AI agents without any coding required.

With Pearl, you can easily participate in Olas Predict: Olas’ most prominent AI agent economy. This economy centres around prediction markets, where AI agents bet on the likelihood of events occurring.

Source: Autonolas

On Pearl, your AI agent assumes the role of the trading agent in the diagram above. The Trader asks the Mech agent to estimate the probability of a prediction market coming true. With that information, the Trader executes an action accordingly.

You can check this out for the full example. After the process is complete, the agent will receive its Olas emissions.

Olas aims to promote the growth and sustainability of agent economies by incentivizing entities that own/create agents through Olas emissions. To qualify for emissions, agent operators must stake OLAS. All of this can be interacted with through Pearl.

Source: Autonolas

P.S. If you’re curious about what’s happening on-chain, you can follow the trail from the wallet address in the Pearl desktop app. Exploring the address on Gnosis Scan provides a first-hand experience with the Olas Open Autonomy Framework and Protocol

Source: Gnosis Scan, an example of a contract that deploys agents. Our Pearl wallet had a few interactions with this contract.

Olas Staking

Olas employs a staking mechanism called Proof of Active Agent to help grow its ecosystem.

Launchers can set specific KPIs and commission autonomous services aligned with their objectives by creating staking contracts.

The DAO plays a key role by voting to direct OLAS emissions toward agent use cases they believe are valuable, providing incentives for builders and operators. Agent operators that meet the defined KPIs are rewarded through staking, with all verifications handled via smart contracts.

So far, entities like GnosisDAO with its Omen prediction markets and OP SuperChain with Optimus have launched agents with emissions directed towards them. Any group or organization that wants to incorporate agents into their user base can create a staking contract to incentivize agent operation.

The Olas Core Team

Olas was founded by Valory, a group of scientists and developers that publish extensive research in AI agent fields, distributed technology, and blockchains. Olas was the first of their projects.

• David Minarsch - CEO/cofounder, led the team that built the framework for MAS in the DLT space
• David Galindo - CTO/cofounder, associate professor of computer security
• Oaksprout the Tan - CPO/cofounder, notable developer in crypto communities

While Olas was launched via an LBP on Fjord Foundry, Valory raised about $4M from True Ventures, Signature Ventures, and others in 2022.

#2: OLAS Tokenomics

OLAS tokenomics is complex, with multiple different supply and demand drivers. The goal is to build a self-sustaining ecosystem of autonomous services.

Let’s break it down.

Supply-side

The initial allocation for OLAS is:

• 32.65% allocated to founding members of the Olas DAO
• 10% allocated to Valory AG to maintain, run, and further develop Olas
• 10% allocated to the Olas DAO treasury
• 47.35% allocated for incentivizing developers and bonders

The token has a 1B token cap in the first ten years. Subsequently, token inflation will be capped at 2% per annum.

Source: Flipside Crypto

As of October 2024, ~540M OLAS have been minted:

• 46M in circulation
• 240M staked as veOLAS
• 253M in the DAO treasury + buOLAS + Valory wallet

As for future supply, ~460M tokens will be minted over the next 8 years. About 56M tokens will be minted in 2025, representing ~10% inflation of total tokens but 121% inflation of curently circulating tokens (!)

Newly minted tokens are distributed as follows:

• 4% to Builders of AI agents
• 46% to Bonders
• 50% to Agent Operators

At the current price of $1.25, that implies that OLAS has a circulating market cap of $57M, total market cap of $675M, and a FDV of $1.25B.

Note that OLAS’s market cap on Coingecko can appear misleadingly low because it does not account for the veOLAS and DAO tokens.

Demand-Side

Utilities of OLAS:

1. Bonding: OLAS holders can provide liquidity on a DEX like Uniswap and sell their LP tokens to acquire OLAS at a discount. This bonding mechanism is similar to Olympus DAO’s model, locking liquidity for the protocol. Currently, users have little incentive to do this, as the value of the LP tokens is generally higher than the OLAS rewards received. There is $11.1M of protocol-owned liquidity across all chains that has generated $2.4M in revenue from LP fees. The fees go to the DAO.

Rewards for LP Bonding

1. Governance: OLAS holders can lock their tokens to receive veOLAS, which grants voting rights and governance power within the DAO. Voters can decide which staking contracts receive the most emissions.
2. Staking: Agent operators can earn OLAS by staking and operating agents. Different agents require different amounts of OLAS staked. Staking yields are quite attractive, currently offering over 130% APY.

Staking rewards for operating AI agents on Olas

Our simple interpretation of the ideal economic flywheel for OLAS works like this:

1. Developers build useful agents on the OLAS platform.
2. These AI agents are used widely and meaningfully productive, generating income & donations for the protocol. Income is shared between the treasury and developers
3. This leads to increased demand for OLAS from agent operators who need to stake OLAS to run agents.,
4. Higher OLAS price incentivises bonders and increases liquidity within the ecosystem.
5. (2 - 4) will fuel more development on Olas, repeating the cycle.

Our Thoughts on OLAS: Keeping On Our Watchlist

OLAS Price Chart. Source: Coingecko

OLAS was launched in October 2023 and has since gone through a chart pattern very typical of low-float, high-FDV tokens — an initial pump based driven by early attention and then slow bleed out over time.

After peaking at $7.69 in January 2024, it has since dropped 84% to $1.25. Because of the low circulating supply, prices can move quickly upwards, but investors have to be wary of sell pressure as new tokens are released into circulation. There are 18,536 OLAS holders across all chains, a reasonably healthy distribution.

We’re keeping OLAS on our Crypto AI token watchlist and staying patient, while looking out for major positive catalysts.

• OLAS’s tokenomic design presents a challenge. The primary inflows into the ecosystem will come from donations and agent operators, which need to offset the significant inflationary pressure. So far, only $96K of donations (from productive agent work) have been received. Demand will likely remain weak until we see a major uptick in AI agent use cases built on Olas. While the prediction market agent is promising, many more are needed to drive heavy growth.
• There is a substantial allocation of tokens to the founding team (8.13% buOLAS, or 81.2M tokens) and Valory (7.19%, or 71.8M tokens). Combined, these represent 15.32% of the total supply—153M tokens in total. On the bright side, there’s reason to believe a portion of these will be locked again as veOLAS; historically, 45.4% of buOLAS has been relocked into veOLAS.

• There is a limited pool of developers building on Olas. The chart above shows the emissions distributed to developers. Notice that despite available incentives, only two epochs distributed emissions to developers. There is an Olas Dev Academy, which could bring in more talented builders in the coming quarters.
• Although OLAS was first to the market for Crypto AI agent protocols, it risks losing mindshare across its ecosystem—developers, operators, and users alike. With agent competitors like Morpheus and Theoriq emerging, the pressure to stay relevant is high.

What Could Change Our View on OLAS:

1. AI agents explode, experiencing major adoption: With advancements in AI, such as OpenAI’s latest models (o1 ‘strawberry’), the AI agent future is becoming closer to reality. This opens up the design space for AI agents in both crypto and outside of crypto. Narrative-wise, Olas has the opportunity to become the go-to platform for on-chain AI agents, making it an index play for anyone who wants to ride on the agent wave.
2. New major AI agent services emerge on Olas. This will kick-start the growth flywheel we outlined earlier. Given how early the whole AI agent space is, it’s hard to say what this will be right now, but it will be obvious in hindsight, and the user metrics will reflect this. Perhaps Optimus —this is an autonomous service for DeFI asset management across the OP Superchain ecosystem and will be launching soon on Olas.
3. Increased developer engagement: If Olas can attract more talented developers to build compelling AI agents that people want to operate, demand for the token could rise, as users will need a minimum amount of OLAS to stake and operate agents. We will closely watch the total number and type of agent transactions on OLAS for growth inflection points.
4. Revamped tokenomics: Introducing stronger demand drivers in the token design could help stabilize inflation and enhance OLAS’s long-term viability. For example, some form of revenue sharing or buy-back-and-burn using income generated. Or the founding team could explicitly outline how they will use their large token allocations to grow the ecosystem.

Disclaimer：

1. This article is reprinted from [Chain of Thought]. All copyrights belong to the original author [Teng Yan & Josh]. If there are objections to this reprint, please contact the Gate Learn team, and they will handle it promptly.
2. Liability Disclaimer: The views and opinions expressed in this article are solely those of the author and do not constitute any investment advice.
3. Translations of the article into other languages are done by the Gate Learn team. Unless mentioned, copying, distributing, or plagiarizing the translated articles is prohibited.