Fetch.ai / ASI Alliance

What it does

The Artificial Superintelligence Alliance is what happens when three separate crypto-AI projects decide they are stronger together, then one of them leaves in a lawsuit.

Fetch.ai, SingularityNET and Ocean Protocol announced their merger in March 2024 under the ASI Alliance banner, unifying under the FET tokenTokenA digital unit of value or access rights tracked on a blockchain. Tokens can represent ownership in a project, a right to use a service, a share of future revenue, or simply a tradable asset with no underlying claim.Like a physical poker chip a casino issues. The chip itself has no value. What makes it worth something is what it lets you do at the casino, what the casino has promised, and how much other people will pay you for it.Token in the glossary →. CUDOS joined in October 2024, contributing 30 validators and compute infrastructure. Ocean Protocol withdrew in October 2025, alleging governance failures, rushed token migration steps and strategic misalignment. A lawsuit followed in November 2025, alleging that Ocean’s foundation had converted 661 million OCEAN tokens into 286 million FET and sold roughly 263 million FET (over 10% of circulating supplyCirculating SupplyThe number of tokens currently in circulation and tradeable on the open market. Differs from total supply (which includes locked or unvested tokens) and max supply (the upper limit, if there is one).Like the number of cars on the road today versus the number ever produced. Some are in showrooms, some in junkyards, some still at the factory. Only the ones on the road count toward what people are actually driving.Circulating Supply in the glossary →) on the open market without disclosure. Media reported a settlement in late 2025, but the case, Fetch Compute, Inc. v. Pon (S.D.N.Y. 1:25-cv-09210), carries no termination on its docket as of September 2026. The plaintiffs are Fetch Compute and three individuals, and the 286 million FET is the conversion the complaint alleges, not an amount returned.

What remains is a federated alliance of three entities: Fetch.ai (autonomous AI agents), SingularityNET (AI services marketplace), and CUDOS (cloud compute). The core technology is Fetch.ai’s Cosmos SDKSDKSoftware Development Kit. A collection of code libraries, documentation, and tools that lets developers integrate a service into their applications without writing everything from scratch. SDKs are how projects become easy to build with.Like a plug-and-play kit for building furniture. You don't have to mill your own wood, forge your own screws, or design the joinery from scratch. The kit gives you pre-cut parts and instructions so you can assemble the thing in an afternoon.SDK in the glossary → blockchain (Fetchhub-4) running 71 bonded validators (the on-chain cap is 91) with CometBFT consensus. The uAgents Python library lets developers build autonomous agents that register on-chain via the Almanac contract. Agentverse provides a no-code browser-based deployment platform. SingularityNET offers a marketplace for publishing and consuming AI services across Ethereum and Cardano. The alliance is also building ASI Chain, a new blockDAG Layer 1 using MeTTa (a cognitive programming language from the OpenCog Hyperon AGI stack) compiled into Rholang. ASI Chain entered DevNet in November 2025.

The leadership: Humayun Sheikh (Fetch.ai founder, DeepMind founding investor) is Chairman of the Governing Council. Ben Goertzel (SingularityNET founder, credited with coining the term “Artificial General Intelligence,” over 150 research publications) is CEO. Together they drive strategy for an alliance that claims “300-plus experts” but publishes no verifiable headcount.

The entity structure is complex. Fetch.ai Foundation (co-founded with Bosch), SingularityNET Foundation (Netherlands), and CUDOS (UK) each remain independent legal entities within the alliance. Total disclosed funding: $61.9 million, including a $40 million Series C led by DWF Labs in March 2023 at a $250 million valuation. The original FET launched via Binance Launchpad IEO in February 2019 ($6 million raised in 10 seconds).

Value proposition

Three layers, each at a different stage of maturity.

Start with the agent layer, which is the most tangible. uAgents (1,640 GitHub stars in a 24 September 2026 read, Apache 2.0) is a functional Python framework for building autonomous AI agents. Agents register on-chain, discover each other through the Almanac contract, and transact in FET. Agentverse removes the coding barrier with one-click deployment templates. This is the “AI agents on blockchain” thesis in its most accessible form, and Fetch.ai has a first-mover advantage (uAgents launched in 2022).

The AI services marketplace is older but less proven. SingularityNET has been offering a marketplace for AI services since 2017. Developers can publish and consume AI models. The OpenCog Hyperon AGI stack is Goertzel’s long-running open-source AGI research project. Ambitious in scope. Usage volumes are not publicly disclosed, which is a problem for a marketplace that has been live for years.

Compute and inferenceInferenceRunning a trained AI model to produce an answer. Inference is what happens when you type a prompt into ChatGPT and get a response. The model takes your input, computes a best guess, and returns it.Like asking an expert for their opinion. The training was the decades they spent becoming an expert. The inference is the 30 seconds it takes them to answer your specific question.Inference in the glossary → is the newest layer. ASI Cloud exited beta in December 2025, claiming prices roughly 50% cheaper than AWS. ASI-1 Mini is positioned as the “first Web3-native LLMLLMLarge Language Model. A neural network trained on vast amounts of text to predict the next word in a sequence. Modern LLMs (GPT, Claude, Llama, Qwen, DeepSeek) generate human-quality text and are the foundation of most modern AI products.Like an autocomplete that read every book ever written. It has no memory of individual texts but it has absorbed the patterns of language so deeply that it can generate paragraphs that sound human. The skill is statistical, not conscious.LLM in the glossary →,” a modelModelA trained neural network that takes inputs (text, images, audio) and produces outputs (more text, classifications, generated content). In DeAI the model is the thing that actually does the work.Like a very experienced apprentice who has spent years watching thousands of masters make furniture. They can't explain how they know when a joint is right, but they can make a chair that looks and functions like a Chippendale. The training is invisible. The output is what matters.Model in the glossary → with four reasoning modes powered by FET payments. Neither product publishes usage metrics.

On 27 April 2026, the Alliance launched OmegaClaw, an open-source neural-symbolic agent toolkit powered by Hyperon, with the first agent (“Oma”) shipping on Telegram. The Alliance describes this as the “final integration phase” of Fetch.ai into the broader ASI:Chain stack. OmegaClaw’s code is open, which is a positive on the Open Source dimension; usage and adoption data is not yet available.

The counter-narrative is simple. Strip away the “Artificial Superintelligence” branding and you have a collection of AI-adjacent tools at different stages of development, operated by a federated alliance that has already lost one member in a lawsuit. The paid products (ASI:One, ASI Cloud, ASI-1 Mini) are centrally operated by Fetch.ai Inc. The decentralisation is at the base layerL1Layer 1. A base blockchain that runs its own consensus mechanism, executes transactions, and settles its own state. Bitcoin, Ethereum, NEAR, and Solana are all L1s. Anything built on top of an L1 is technically a Layer 2 or higher.Like the foundation of a building. Nothing else can exist on top until the foundation is solid. Different L1s make different tradeoffs for what kind of building they can support.L1 in the glossary → (Cosmos chain, stakingStakingLocking up a cryptocurrency to help secure a blockchain network, usually in exchange for rewards. The locked tokens act as a security deposit that can be taken away if the staker misbehaves.Like putting down a large rental deposit for an apartment. You get the money back if you behave, you earn interest while it's locked, and the landlord takes it if you trash the place.Staking in the glossary →, governance). The application layer is a company running servers.

For the sovereignty thesis: the open-source agent runtime and the permissionless base layer have substantive merit. But the core AI products are centralised services. You are trusting Fetch.ai Inc. with your inference workloads, not running them on decentralised infrastructure.

Tokenomics

FET tokenomics turn on one gap: FET pays for staking, gas and agent registration on Fetchhub-4 and for ASI:One and ASI Cloud usage, but Fetch.ai publishes no revenue for any product, so demand for the token is unmeasured. Value reaches holders only through burns, and the record is thin: the announced Earn and Burn plan has produced two discretionary burns from one Safe on chain, 5 million FET on 10 January 2025 and 109,350 FET on 5 November 2025, with no contract linking them to product fees and none since. A $50 million buyback announced in June 2025 has no execution evidence. Against that, Fetchhub-4 mints new FET at 3% a year, about 53 million FET, paid to stakers, and founder and future-release allocations keep vesting to 2050 (full breakdown).

FET is one of the most complex tokens in the space because it absorbed three other token communities.

The merger converted existing holders at fixed ratios: 1 AGIX = 0.433350 FET, 1 OCEAN = 0.433226 FET, 1 CUDOS at 118.344:1 (with a 5% fee and 3-10 month vestingVestingA schedule that locks up tokens allocated to insiders, investors, and team members, releasing them gradually over months or years. Vesting prevents insiders from dumping on public buyers immediately after launch.Like a new employee's stock options at a startup. You don't get all the shares on day one. They unlock over four years so you stick around and do the work rather than cashing out and leaving.Vesting in the glossary →). Total supply post-merger: 2.71 billion FET. Circulating supply: 2.3 billion (85.2%). The full unlock schedule extends to 2050.

Distribution of the merged supply:

• AGIX migration (SingularityNET holders): 32.95%
• OCEAN migration (Ocean Protocol holders): 23.22%
• Founders: 8.77% (cliffCliffA waiting period at the start of a token vesting schedule during which no tokens unlock at all. After the cliff ends, tokens begin releasing according to the vesting schedule.Like a probationary period at a new job. You don't get your stock options on day one. You wait 12 months to prove you'll stick around, then everything starts unlocking normally.Cliff in the glossary → vesting to 2050)
• Foundation: 8.77% (Fetch.ai Foundation, co-founded with Bosch)
• Future releases: 7.63% (reserved, schedule to 2050)
• MiningProof of WorkThe original blockchain consensus mechanism where miners compete to solve computationally expensive puzzles. The winner proposes the next block and earns the rewards. Proof of Work secures Bitcoin and most pre-2020 chains.Like a lottery that runs every 10 minutes where the tickets cost electricity. Whoever spends the most electricity buying lottery tickets has the best chance of winning that round's prize. Nobody can fake the result because the proof of their work is verifiable by everyone.Proof of Work in the glossary →/staking rewards: 6.57%
• Private sale: 5.08% (fully unlocked)
• Advisors: 4.38% (cliff vesting)
• Public sale (Binance Launchpad IEO): 2.49%

The AGIX and OCEAN migration allocations (56.17% combined) are not new insider supply; they represent conversion of existing communities. But the roughly 600 million FET minted to absorb OCEAN holders did dilute existing FET holders. Ocean Protocol sold approximately 263 million of those FET on the open market, leading to a lawsuit, still live in SDNY, over an alleged 286 million FET conversion.

The ASI Alliance describes Earn and Burn as ecosystem fees from ASI Cloud, ASI-1 Mini and other products used to buy and burnBurnPermanently removing tokens from circulation by sending them to an address that no one controls. Burns reduce total supply, which (all else equal) makes each remaining token worth more of the network's value.Like a company buying back its own shares and shredding them. The company's total value stays the same, but each remaining share now represents a slightly bigger slice of that value.Burn in the glossary → FET. On chain, both burns were burn() calls from the same Safe, and the source of the burned FET isn’t traced. The first burn was 0.18% of supply and the second 0.004%, so the programme is symbolic. The $50 million FET buybackBuybackUsing protocol revenue to purchase tokens on the open market, usually to burn them or return them to a treasury. Buybacks convert business income into upward pressure on the token by reducing circulating supply.Like a public company using profits to repurchase and retire its own shares. The cash leaves the company's balance sheet, the share count drops, and every remaining shareholder owns a slightly bigger slice of the same business.Buyback in the glossary → announced in June 2025, to run across several exchanges, has no completion report or on-chain trail.

Staking yields 5.5-8% APY with a standard Cosmos 21-day unbonding period. That yield comes mainly from the 3% mint shared among bonded stakers, so it redistributes new supply rather than paying out fee income. About 605 million FET was bonded on Fetchhub-4 in a 24 September 2026 chain read, roughly a quarter of circulating supply. Standard Cosmos slashing: 5% for double-signing, 0.01% for downtime.

FET is down more than 90% from the all-time high of $3.45 from March 2024. Listed on Binance, Coinbase, Kraken, OKX, Bybit and others. See live data above for current pricing.

How to participate

Stake FET. Delegate to validators on Fetchhub-4. Earn 5.5-8% APY. Use the ASI Alliance WalletWalletSoftware that stores the private keys needed to control tokens on a blockchain. A wallet does not actually hold any tokens. The tokens live on the chain. The wallet holds the keys that prove you own them.Like the key to a safe deposit box. The key doesn't contain your valuables. The valuables sit in the bank's vault. The key is what proves you're allowed to open the box and take them.Wallet in the glossary → (browser extension) or any Cosmos-compatible wallet (Keplr, Cosmostation). No minimum stake; keep at least 1 FET undelegated for transaction fees. Twenty-one-day unbonding period. Choose validators carefully; top validators (Kiln at about 122 million FET, Colossus at 111 million, Binance Staking at 82 million in a 24 September 2026 read) hold disproportionate stake.

Build agents. The most interesting participation mode. Write autonomous AI agents in Python using the uAgents SDK. Deploy on Agentverse (no-code, browser-based) or self-host. Register agents on the Almanac contract (requires FET, renewed every 48 hours). Access ASI-1 Mini LLM for agent intelligence. Requires Python skills and understanding of the agent architecture.

Use products. ASI Cloud for GPUGPUGraphics Processing Unit. Originally designed to render video game graphics, GPUs turned out to be exceptionally good at the massively parallel math that AI models need. Modern AI training and inference runs almost entirely on GPUs.Like a factory with 10,000 workers doing the same simple task in parallel, versus a CPU which is more like 10 workers each doing different complex tasks. AI training involves doing simple math a million times per second on a million numbers, which is exactly what the GPU factory is designed for.GPU in the glossary → inference (reportedly 50% cheaper than AWS, but unverified). ASI-1 Mini for LLM access via APIAPIApplication Programming Interface. A structured way for one piece of software to talk to another. In DeAI, APIs let applications request inference from a model without running the model themselves.Like a waiter in a restaurant. You don't walk into the kitchen and cook your own meal. You tell the waiter what you want, they tell the kitchen, the kitchen cooks it, and the waiter brings it back. The API is the waiter.API in the glossary →. Browse and consume AI services on the SingularityNET marketplace. Pay in FET.

Validate. Run a validator node on Fetchhub-4. The chain allows 91 slots and 71 were bonded in September 2026. Requires substantial FET stake (competitive; top validators hold 30 million-plus FET), server infrastructure and DevOps expertise. High barrier to entry.

Governance. Vote on alliance proposals using staked FET. Proposals have included the CUDOS merger approval and tokenomics changes. A proposed Paal.ai merger (900 million PAAL tokens converting to FET at 6.24:1 with 180-day vesting) was withdrawn in November 2024 after community backlash over the conversion ratio, causing PAAL to drop 70%.

Contribute. Open-source repos across three GitHub organisations (fetchai, singnet, oceanprotocol). Innovation Lab partnership with Imperial College London backed by a $10 million fund. Hackathon participation. Apache 2.0 licensing on core repos.

Honest assessment

What works

Start with uAgents, the most mature open-source agent framework in the DeAI space. Functional, well-documented, Apache 2.0, actively maintained (pushed 24 September 2026). Fetchhub-4 runs reliably as a Cosmos SDK chain with IBC interoperability, and 375 public repositories across three GitHub organisations in a September 2026 read is a substantial open-source footprint.

Team credibility is substantive. Goertzel is a credible AGI research figure with 150-plus publications. Sheikh is a DeepMind founding investor. The Innovation Lab with Imperial College London, backed by a $10 million fund, is a serious academic partnership, not a PR announcement. The Earn and Burn plan has the right shape, though its two burns to date were discretionary and none has run since November 2025. ASI-1 Mini, whatever you think of the “first Web3-native LLM” framing, is a concrete product at the intersection of AI and crypto.

What does not work

The merger has been a governance catastrophe. Ocean Protocol’s exit, the sale of 263 million FET on the open market, and the resulting lawsuit, still unresolved, are not growing pains. They are structural failures in alliance governance. The “Artificial Superintelligence” branding is aspirational to the point of misleading. The constituent projects are building AI agents, marketplaces and compute infrastructure. Valuable, yes. A path to superintelligence? No.

Usage metrics are notably opaque. No public dashboard for transaction counts, active users or revenue. ASI Cloud exited beta in December 2025 but publishes no adoption data. SingularityNET’s marketplace has been live since 2017 and still doesn’t disclose usage volumes. For a project of this market cap, this lack of transparency is not a gap. It’s a choice.

ASI Chain (blockDAG with MeTTa/Rholang) is in DevNet with no mainnet timeline. The complexity of running three separate technical platforms (Cosmos chain, Ethereum/Cardano marketplace, DevNet blockDAG) under one token creates integration challenges and a confusing user experience.

The risk

Bridge custody is the risk that fired. On 19 September 2026 at 20:21 UTC an attacker called conversionIn on TokenConversionManagerV3 with a valid signature and took 8,721,530.40 FET, the entire remaining escrow in that contract, whose balance fell from 8.72 million to zero in one transaction. FET supply is untouched at 2,714,384,547, so this was a drain of tokens held for cross-chain conversion rather than an issuance event, and it cost roughly $1.55 million. The same receiving address minted 408.5 million NTX on NuNet’s token twenty-nine minutes later using a separately compromised key, which is what makes this one attacker rather than two. The token contract itself, behind a 3-of-5 Safe, was never involved. The conversion path beside it authorised releases on one off-chain signature.

SingularityNET says an unauthorised party reached part of its cloud infrastructure and used it to mint tokens and withdraw assets through its bridges, which it has since deactivated. The AGIX contract shows the scale: in the early hours of 20 September (UTC) its TokenConversionManager minted 895,962,344.71 AGIX in 90 conversionIn calls to two addresses, lifting AGIX total supply from about 208.1 million to 1,104.0 million. SingularityNET has paused AGIX on Ethereum and its AGIX migration, and on 24 September said the legacy token will be retired for a replacement issued to affected holders. It reports that FET minting controls weren’t affected, which matches the unchanged FET supply above.

Alliance fragmentation is the defining risk. Ocean Protocol already left. If SingularityNET departed, the alliance narrative collapses entirely. The Ocean lawsuit is still live in SDNY, so the legal overhang remains, and the episode exposed governance vulnerabilities that remain unaddressed.

Down more than 90% from the all-time high. The merger was announced at a combined “valuation” of $7.5 billion. That gap between hype valuation and current reality represents real losses for holders who bought the merger narrative.

Key person dependency is acute. Sheikh and Goertzel are essential to the alliance. Goertzel’s broad research interests (AGI, robotics, philosophy, OpenCog Hyperon) raise focus risk, and the federated structure means each foundation retains independent governance, creating potential for further misalignment.

Core paid products (ASI:One, ASI Cloud, ASI-1 Mini, Agentverse) are centrally operated by Fetch.ai Inc. This contradicts the decentralisation narrative. Fetchhub-4’s 71 bonded validators, with the top two together past the one-third that can halt the chain, create concentration at the infrastructure level too.

My position

I don’t hold FET. The governance failures demonstrated by the Ocean Protocol exit, the opaque usage metrics, and the decline of more than 90% from the merger-hype peak make this a wait-and-see proposition. The uAgents SDK is worth exploring as a developer tool. But “wait-and-see” here means: show me a public revenue dashboard, show me governance that doesn’t end in lawsuits, then we’ll talk.

Freedom Score: 49/100

The ASI Alliance scores 49/100 (D grade). Full methodology at Freedom Score Methodology.

Infrastructure decentralisation (9/20): Fetchhub-4 runs on Cosmos SDK with CometBFT consensus, with 71 validators bonded against an on-chain cap of 91. Twenty slots sit empty, so the barrier to entry is capital. Stake is concentrated: in a 25 September 2026 read Kiln held 20.0% and Colossus 18.3%, together past the one-third that can halt the chain, and the top five held 67.4%. The 30 CUDOS validators added some diversity.

The application layer is more centralised than the chain. Core value-adding products (Agentverse, ASI Cloud, ASI-1 Mini) are centrally hosted by Fetch.ai Inc, and the SingularityNET marketplace relies on centralised infrastructure for service discovery. ASI Chain is still in DevNet.

Governance decentralisation (8/20): On-chain governanceDAODecentralised Autonomous Organisation. A way to coordinate decisions and manage a treasury using token-weighted voting instead of a traditional company structure. Token holders propose and vote on changes directly.Like a shareholder-run company where every shareholder can vote on every decision, the votes are public, and the company can't do anything the shareholders don't approve. The coordination is messier than a normal company but nobody has unilateral control.DAO in the glossary → via FET token voting exists for major decisions (merger approvals, new members). The CUDOS merger went through a community vote; the Paal.ai proposal was withdrawn before voting after community backlash. However, the Governing Council (Sheikh as Chairman, Goertzel as CEO) holds outsized influence. Ocean Protocol’s exit exposed governance failures: Ocean alleged the governance “didn’t match the original decentralised vision” and cited rushed decision-making. Each foundation retains independent governance, enabling strategic decisions without full community consent. The lawsuit, still unresolved, demonstrates governance dysfunction.

Token distribution fairness (7/15): Original FET launched via Binance Launchpad IEO ($6M raised, February 2019). Original insider allocations (founders 8.77%, foundation 8.77%, advisors 4.38%, private sale 5.08%) total roughly 27%. The AGIX and OCEAN migration allocations (56.17% combined) represent community conversion, not new insider supply. However, a still-live SDNY suit alleges that Ocean Foundation converted 661M OCEAN into 286M FET and sold 263M FET without disclosure. CUDOS merger included a 5% fee. The $50M buyback announced in June 2025 has no execution evidence, so it earns no credit here.

Censorship resistance (7/15): The Cosmos SDK base layer is permissionless, and IBC enables cross-chain interoperability. The Ethereum token is itself a censorship vector: the FetchToken ERC-20 (0xaea46a60368a7bd060eec7df8cba43b7ef41ad85) has a working pause() that freezes every transfer at once, and its pauser and minter roles sit with one 3-of-5 Safe. On Fetchhub-4, two validators hold enough stake to halt block production.

The application layer has further vectors. Agentverse is centrally hosted (Fetch.ai can deplatform agents), ASI Cloud and ASI-1 Mini are centrally operated, and the SingularityNET marketplace has curation capabilities. Agent registration requires renewal every 48 hours via the Almanac contract, creating dependency.

Data sovereignty (8/15): Ocean Protocol’s compute-to-data paradigm was the alliance’s strongest data sovereignty feature, enabling computation on data without the data leaving the provider’s infrastructure. Ocean Protocol withdrew in October 2025, removing this capability. The uAgents runtime gives developers control over agent logic and data. However, ASI Cloud and ASI-1 Mini are centrally hosted; user queries pass through Fetch.ai’s infrastructure. No ZKZKZero Knowledge. A class of cryptographic proofs that let you prove something is true without revealing any of the underlying information. ZK lets a network verify a transaction without seeing the transaction's contents.Like proving you know the password to a safe by demonstrating you can open it, without ever saying the password out loud. The verifier learns that you know the password and nothing more.ZK in the glossary → proofs or advanced privacy-preserving features in the core stack.

Open source and transparency (10/15): Strong open-source foundation: 375 public repos across three GitHub organisations in a September 2026 read (fetchai: 96, singnet: 182, oceanprotocol: 97). Core infrastructure (uAgents, fetchd, cosmpy, SingularityNET daemon) is Apache 2.0. Active development continues. However, ASI-1 Mini model weightsParametersThe internal numbers (weights and biases) inside a neural network that get adjusted during training. A 70-billion-parameter model has 70 billion adjustable internal numbers encoding everything it has learned.Like the synapses in a human brain. Each parameter is a tiny dial that gets nudged a little during training. With enough dials, the network can represent surprisingly complex patterns. The total parameter count is roughly how much "brain" the model has.Parameters in the glossary → are not confirmed open source, ASI Cloud infrastructure is not open, and ASI Chain implementation details are limited. The Ocean exit revealed a lack of transparency in token operations. Usage metrics are not publicly reported.

Fetch’s own node software slipped in September. Proposal 41 passed on 16 September with no votes against and moved Fetchhub-4 to fetchd v0.15.1 at height 29,254,333. Fetch shipped it as a binary-only release, and its upgrade guide says it is “not possible to build the node binary from the publicly accessible source code”. The release pins patched wasmd and wasmvm forks in Fetch’s GitHub organisation, and by 4 October those repositories and the pinned commits were public, but validators upgraded on a binary they could not build themselves.

Path to improvement

Three changes would materially increase the ASI Alliance’s score:

1. Publish transparent usage metrics. No public dashboard for ASI Cloud adoption, SingularityNET marketplace volumes, agent deployments or revenue. This is indefensible for a project at this scale. A real-time metrics dashboard would be the single biggest credibility signal. Akash publishes quarterly Messari reports and tracks on-chain revenue. The ASI Alliance publishes nothing.
2. Publish a post-mortem and reform governance. The Ocean lawsuit is still unresolved, and the alliance hasn’t published a post-mortem or implemented governance reforms. Clear exit procedures, transparent treasury operations and community veto mechanisms would address the structural failures the episode exposed.
3. Open-source ASI-1 Mini and ASI Cloud. The core AI products are centrally operated and not open source. Open-sourcing ASI-1 Mini model weights and the ASI Cloud orchestration layer would align the products with the open-source credentials of the underlying infrastructure. The “decentralised AIDeAIDecentralised AI. An umbrella term for blockchain-based projects that build AI infrastructure (compute, data, inference, models, agents) without a single central provider controlling the system.Like the difference between streaming a movie from Netflix and sharing it via BitTorrent. Netflix is fast and polished but one company controls what you can watch and what you pay. BitTorrent is messier but no single operator can shut you out.DeAI in the glossary →” narrative can’t hold when the AI itself is proprietary.

Returns Score: 55/100

FET scores 55/100 (C grade). Full methodology at Returns Score Methodology.

Token utility (14/20): FET is the staking token for Fetchhub-4 validators, the governance token for alliance-wide decisions, and the payment token for agent services and ASI Cloud compute. Those are three working utility layers, which is more than most projects in this space can claim. The complication is the merger. Three separate communities (Fetch.ai, SingularityNET, and CUDOS) were funnelled into one token, and the AGIX and CUDOS migration ratios created confusion about what FET actually represents. The utility functions but the identity is muddled.

Value accrual (8/20): The merger has made value accrual mechanics hard to evaluate, and the evidence so far isn’t encouraging. Ocean Protocol’s withdrawal and sale of 263 million FET on the open market demonstrated that the alliance structure can actively destroy value rather than create it. Fetch Compute’s lawsuit over the 286 million FET is still open, so none of that damage has been unwound, and the episode set a precedent. ASI Cloud and ASI-1 Mini charge for usage, but both are centrally operated by Fetch.ai Inc. The fees flow to a company, and no contract routes any of them to FET holders. The two on-chain burns (5 million FET in January 2025, 109,350 FET in November 2025) were discretionary, and the $50 million buyback announced in June 2025 has no execution evidence. Neither is a sustainable accrual mechanism.

Supply dynamics (10/20): The 2.71 billion total absorbed three separate token supplies via migration, which is inherently dilutive to original FET holders who did not sign up for the additional supply. Ocean Protocol sold 263 million FET (over 10% of circulating supply) without disclosure, a supply shock of extraordinary magnitude. No FET has been returned, and the sell pressure already hit the market. CUDOS migration included a 5% fee. The structural risk here is the precedent: alliance membership can introduce surprise supply events. The total is also not capped in code: a 3-of-5 Safe holds an uncapped mint, so the 2.71 billion figure rests on multisig policy. Fetchhub-4 separately mints new FET at 3% a year to stakers, which the 2.71 billion ERC-20 total doesn’t show.

Revenue sustainability (10/25): The uAgents SDK is functional and has first-mover advantage in the agent space. Agentverse removes the developer barrier. ASI Cloud claims 50% cheaper than AWS. ASI-1 Mini offers inference with FET payments. The product breadth is wide: multiple revenue-generating surfaces exist, and the toolkit has been live long enough to accumulate a developer base. The problem is that none of these products publish usage metrics. For a project that has been building since 2017, the absence of public revenue data, marketplace volumes, or active user counts is a deliberate choice, and not one that inspires confidence. Fetch.ai publishes no revenue figure for ASI:One, ASI Cloud or Agentverse, and neither Token Terminal nor DeFiLlama tracks one, so the market will price it conservatively.

LiquidityLiquidityHow easily a token can be bought or sold without moving the price. High liquidity means you can enter or exit large positions quickly at the quoted price. Low liquidity means even small trades can swing the market.Like the difference between selling a house and selling a share of Apple stock. The house might be worth more on paper, but finding a buyer at that price takes weeks. The Apple share converts to cash in one click.Liquidity in the glossary → and access (13/15): This is FET’s strongest dimension. Major CEX listings across Binance, Coinbase, Kraken, and others provide institutional-grade liquidity. Strong daily volume means you can enter and exit meaningful positions without moving the market. The Binance Launchpad IEO origin gives the token deep exchange relationships. For all the project’s problems, you can at least trade it properly.

Path to improvement

Three changes would materially increase the ASI Alliance’s returns score:

1. Publish auditable revenue metrics for all products. ASI Cloud, ASI-1 Mini, Agentverse, and SingularityNET marketplace should each have public dashboards showing active users, transaction volumes, and revenue. The market can’t value what it can’t measure, and the persistent opacity is being priced as a risk premium.
2. Establish a transparent, recurring buyback-and-burn funded by platform revenue. The $50 million buyback was announced as a one-off and has no execution evidence. A monthly or quarterly burn funded by a fixed percentage of ASI Cloud and marketplace fees would create predictable deflationary pressure and demonstrate that product revenue reaches token holders rather than stopping at Fetch.ai Inc’s balance sheet.
3. Publish a governance post-mortem. The Ocean lawsuit is still live, so the legal overhang remains. The alliance also hasn’t published a transparent account of what happened or what governance reforms will prevent a recurrence. Without that, the governance risk discount persists.

Score change log

Date	Score	Change	Reason
2026-10-06	Data	N/A	The first Earn and Burn burn was dated January 2026; the chain shows 10 January 2025, plus a second burn of 109,350 FET on 5 November 2025. The $50M buyback is now described as announced, not executed. Scores unchanged.
2026-10-06	Data	N/A	Tokenomics now opens on what FET is for, how value reaches holders, what is live, and the 3% Fetchhub-4 mint to stakers. Scores unchanged; Supply Dynamics flagged for review against the mint.
2026-10-04	Data	N/A	Added the SingularityNET breach follow-up: 895,962,344.71 AGIX minted through its conversion manager on 20 September (read on chain), AGIX paused and to be replaced. Scores unchanged.
2026-10-04	Data	N/A	Open source rationale notes proposal 41’s binary-only fetchd v0.15.1 upgrade, whose patched dependencies are now public on GitHub. Scores unchanged.
2026-09-25	Score	Freedom 50 → 49 (D held)	Censorship Resistance 8 → 7: the Ethereum token was credited as adding resistance, but FetchToken has a working pause() held by a 3-of-5 Safe. Validator figures refreshed (71 bonded of 91; two can halt the chain).
2026-09-25	Data	N/A	Six “settled” and “286M FET returned” references the 24 September correction missed (frontmatter, a con, the stat card, two rationale paragraphs) now say the case is open. Scores unchanged.
2026-09-24	Data	N/A	The Ocean lawsuit was described as settled with 286M FET returned in 13 places. Fetch Compute, Inc. v. Pon (SDNY) shows no termination and the 286M FET is an alleged conversion; corrected throughout. Scores unchanged.
2026-09-24	Data	N/A	Phase 5 sweep: circulating supply bound live, validator figures refreshed from a Fetchhub-4 read (70 bonded of a 91 cap, Kiln about 122M FET, ~605M bonded), star, repo and drawdown figures refreshed, and the Supply Dynamics text now says the 2.71B total has no cap in code. Scores unchanged.
2026-09-23	Data	N/A	Supply type corrected to uncapped: the verified FET contract’s mint has no cap, and the ASI page states a total, not a maximum. Scores unchanged.
2026-09-23	Data	N/A	Integrity records re-cut: 1 claim now quotes the ASI Alliance’s own pages verbatim and 6 are our own measurements. Scores unchanged.
2026-09-20	Data	N/A	TokenConversionManagerV3 drained of its entire 8,721,530.40 FET escrow on 19 September via a validly signed conversionIn; FET supply unchanged, so the loss is escrow, not issuance. Scores held; the single-signer conversion path is flagged for the October review.
2026-09-06	Data	N/A	WS3b distribution rail: on-chain top-10 verdict added. 164,757 holders, the widest base in the corpus, with raw top-10 44.98% falling to 30.86% infra-excluded and the largest position a Gnosis Safe at 10.23%. Token Distribution Fairness 7/15 held.
2026-08-12	Data	N/A	WS3 on-chain reconciliation (1->4 verdicts). Supply matches but the cap is NOT code-enforced: FetchToken exposes an uncapped mint() and a global pause(), all roles held by one un-renounced 3-of-5 Gnosis Safe (established). Open-source confirmed (uAgents/fetchd/snet-cli permissive, active; match). Scores unchanged.
2026-05-02	Editorial	N/A	Added OmegaClaw open-source neural-symbolic agent framework launch (27 April 2026, first agent “Oma” on Telegram, framed as final integration phase of Fetch.ai into ASI:Chain). No score change.
2026-04-06	Data	N/A	Quarterly review. Ocean lawsuit settlement and FET return verified. Scores unchanged.