# MOR vs TAO vs FET: Token Models Compared

> Bittensor vs Fetch vs Morpheus: how TAO, FET/ASI and MOR compare on insider allocation, value accrual and actual decentralisation.

- Published: 2026-03-02 · Updated: 2026-10-20
- Canonical: https://ownyourmind.ai/tokenomics/mor-vs-tao-vs-fet/
- Cite as: Own Your Mind (ownyourmind.ai), licence CC BY 4.0. Link the canonical URL.

Charts and interactive widgets are omitted here; project data is at https://ownyourmind.ai/api/scores.json.

---

## Why this comparison matters

MOR (Morpheus), TAO (Bittensor) and FET (Fetch.ai / ASI Alliance) represent three fundamentally different approaches to tokenising AI infrastructure. They look similar from the outside. All three are "AI tokens." In practice, their economic models, incentive structures and decentralisation properties are different enough that comparing them reveals what actually matters in DeAI tokenomics.

## The three models at a glance

## MOR: the [fair-launch compute token](https://ownyourmind.ai/tokenomics/how-mor-actually-works/)

**Model.** Participants contribute capital (stETH, USDC, USDT or WBTC), compute, code or community effort. In return they receive daily MOR emissions proportional to their contribution. Users hold MOR to access compute on the network via an inference-per-second (IPS) quota: the more MOR held, the more compute available. Separate from usage, capital yield buys MOR, half of which is permanently burned, with the other half locked for tail emissions (MRC43); the protocol-owned liquidity leg stopped in January 2025.

**What drives value.** Compute access demand and structural burn. MOR must be staked to access inference on the network and is returned when the stake ends, so demand to hold it scales with usage and falls when usage does. The yield-funded burn reduces supply independently of compute queries. If demand for holding MOR grows while burns reduce supply, the token appreciates.

**Strengths.** Zero insider allocation means no insider selling pressure. The emission model is transparent and predictable. All four contributor groups are incentivised. The Power Factor multiplier, capped near 10.7x, rewards long-term conviction, though it is anchored to a fixed 2024 to 2040 window so the same lock is worth less the later you start it. Capital deposits are withdrawable after just 7 days, and earned MOR has no vesting or lock (the Power Factor lock is entirely voluntary). Capital yield keeps buying MOR to burn: 21 generation events deployed roughly 290 wstETH of yield in the year to 15 June 2026 (Arbiscan). Protocol-owned liquidity itself has not grown since January 2025, so the buying now feeds the burn and the tail-emission lock only.

**Weaknesses.** The network is still young, so compute demand hasn't yet reached the levels needed for significant burn pressure. Capital provider returns depend heavily on MOR price, which is volatile. Liquidity is thin (primarily DEX-based, with low daily volume), making meaningful positions difficult to enter or exit.

**Alignment score.** High. The tokenomics directly incentivise the behaviours the network needs: compute provision, code contribution, capital commitment and community growth. The caveat is control: a 5-of-9 Safe holds owner and upgrade authority over the core contracts, with no timelock.

## TAO: the competitive intelligence token

**Model.** Miners compete within subnets to produce the highest-quality AI outputs. Validators assess quality and allocate rewards. Since [dTAO](https://ownyourmind.ai/tokenomics/bittensor-dtao-subnet-economics/) (February 2025), emission allocation across subnets is set by the market price of each subnet's alpha token, smoothed as a moving average since June 2026, rather than by Root Network votes. Staking into a subnet converts TAO to that subnet's alpha token via an AMM, creating per-subnet price discovery. Emissions split 41% to miners, 41% to validators and 18% to subnet owners by default, and since October 2026 each subnet owner can reset that split, taking up to 33%.

**What drives value.** Subnet quality, staking flows and alpha token dynamics. As subnets attract more stake and their alpha price rises, they earn more emissions. Registration costs, paid in TAO into subnet pools, create friction that regulates network growth. The [first halving](https://ownyourmind.ai/tokenomics/bittensor-halving-what-tao-holders-need-to-know/) completed in December 2025, reducing emissions from ~7,200 TAO/day to ~3,600 TAO/day and strengthening the scarcity narrative.

**Strengths.** The competitive mechanism is innovative. It creates selection pressure that drives improvement. The 21 million cap with Bitcoin-style halving creates a strong scarcity narrative. The subnet architecture allows the network to expand into any AI domain. dTAO moved emission allocation from a centralised Root Network to a market-driven staking mechanism.

**Weaknesses.** Early mining concentration is severe: on the median subnet in our own chain index a single coldkey holds 58.7% of the stake. No formal insider allocation, but 5.38M TAO were mined between January 2021 and October 2023 among a tiny participant group. Polychain Capital incubated the project; DCG and dao5 accumulated large positions early. Governance is still concentrated. The runtime carries no on-chain proposal or vote pallet, and privileged operations including upgrades are dispatched through a single sudo key. The competitive dynamic favours participants with the most capital and best hardware, driving centralisation within the network. On revenue: Pine Analytics (independent Web3 analytics, 24 March 2026) estimates $3-15M in identifiable external revenue across the entire network, one to two orders of magnitude below what the market values it at. The "cheaper than centralised alternatives" pricing in leading subnets is subsidised by TAO emissions, not structural efficiency. (figures as of 2026-10-19)

**Alignment score.** Moderate. The competitive mechanism drives quality but also drives concentration. dTAO improved emission allocation, but stake concentration means the same wealthy participants control most of the economic power. The most successful miners are those with the most resources, which works against the decentralisation thesis over time.

## FET: the enterprise-pivot token

**Model.** FET originated as the Fetch.ai token for autonomous economic agents. In 2024 it merged into the [ASI (Artificial Superintelligence) Alliance](https://ownyourmind.ai/tokenomics/asi-alliance-merged-tokenomics/) with SingularityNET (AGIX), Ocean Protocol (OCEAN) and later CUDOS. The token still trades under the FET ticker; the planned migration to an ASI ticker hasn't completed. In October 2025, Ocean Protocol withdrew from the alliance entirely, and Fetch Compute plus three individuals filed a putative class action (*Fetch Compute, Inc. v. Pon*, S.D.N.Y. 1:25-cv-09210) alleging Ocean converted ~286M FET and sold ~263M FET (~10% of circulating supply). The docket shows no termination, so the case is still live. The alliance is now effectively Fetch.ai + SingularityNET + CUDOS.

**What drives value.** Agent adoption, staking demand and AI services. FET is staked on the Cosmos-based Fetchhub-4 chain for network security (~5.5-8% APY on third-party staking sites) with a 21-day unbonding period. The uAgents framework and Agentverse platform enable autonomous agent development. SingularityNET contributes an AI marketplace, and CUDOS provides compute infrastructure (ASI Cloud).

**Strengths.** Broad ecosystem spanning agents, AI marketplace and compute under one token. Strong marketing and institutional partnerships. High liquidity and exchange availability, making it the most accessible DeAI position for institutional investors. The Fetch.ai chain (Cosmos SDK) is technically sound.

**Weaknesses.** Insider allocations are among the highest in DeAI: roughly 50% of the original FET supply went to founders (20%), foundation (20%) and advisors (10%). The merger added complexity without clear integration of the underlying technology stacks. As of March 2026, Fetch.ai's Cosmos chain, SingularityNET's Ethereum/Cardano marketplace and CUDOS compute still operate as distinct platforms. The planned ASI Chain (a new blockDAG L1) is in DevNet but not yet on mainnet. The "superintelligence" branding is aspirational rather than descriptive. Ocean's departure and the still-unresolved lawsuit exposed governance fractures and the risks of centralised alliance management. The validator set is capped at 91, and just two validators hold enough stake to halt the chain.

**Alignment score.** Low. The token economics are closer to a traditional VC-backed crypto project than a community-owned network. Insider allocations and foundation holdings create structural selling pressure. The governance model concentrates power in the Alliance leadership (Superintelligence Alliance Ltd, a Singapore-registered company). Ocean's exit and the still-live lawsuit demonstrated that "decentralised alliance" governance can fracture when corporate interests diverge.

## The comparison that matters

### Insider allocation risk

More than anything else, this is what determines long-term token behaviour. When insiders hold significant supply, their incentives diverge from other participants. They're incentivised to pump narratives to create exit liquidity, not to build sustainable infrastructure.

- **MOR:** 0% insider allocation. No insider selling pressure.
- **TAO:** No formal insider allocation, but early mining was concentrated among a small group including Polychain, DCG and dao5. On the median subnet in our chain index a single coldkey holds 58.7% of the stake. The effect is similar to insider allocation. (figures as of 2026-10-19)
- **FET/ASI:** ~50% of original FET supply to founders, foundation and advisors. Multiple institutional raises totalling ~$83M. Ocean's alleged sale of ~263M FET, still before the court, shows the selling pressure risk from large holders.

**MOR allocation**

**FET allocation (original)**

### Value accrual mechanism

Where does value flow when the network grows?

### Practical decentralisation

How decentralised is the network in practice, not marketing?

### Bittensor vs Fetch, head to head

TAO and FET are the two most-compared tokens in this set, and they fail the decentralisation test from opposite directions.

Bittensor never sold a token. TAO launched as a pure mining network with a 21 million cap and no formal insider allocation, which reads well until you look at where the stake landed. Our chain index finds a single coldkey holding 58.7% of the stake on the median subnet, and the early mining window (January 2021 to October 2023) concentrated supply among a small group that included Polychain, DCG and dao5. A fair launch, a concentrated outcome. (figures as of 2026-10-19)

Fetch went the other way. The current ASI Alliance token is migration-heavy: SingularityNET and Ocean holders together account for more than half of supply, while the Fetch-native founders, foundation and advisors buckets are roughly a fifth of the token today, down from about half at the original 2019 Fetch.ai launch. Governance runs through a Singapore-registered company, the validator set is capped at 91 with two validators able to halt the chain,  and Ocean's 2025 exit and the still-unresolved lawsuit showed how quickly an alliance can fracture.

On the decentralisation question, Bittensor wins on launch and mechanism: no premine, a hard cap and a competitive design FET has nothing equivalent to. Fetch wins on liquidity and institutional access. Both lose on who controls the network today, TAO to stake concentration and FET to a corporate alliance structure. Morpheus is the only token here with a zero-insider fair launch, which is why it sits alongside them.

## Which to hold

This is not financial advice. It's how I think about these positions.

**MOR** if you believe in the compute network thesis and want exposure to a fair-launched project. Highest alignment between tokenomics and decentralisation goals. Highest risk because the network is youngest.

**TAO** if you want exposure to Bittensor's decentralised AI subnet network and believe the competitive subnet model will drive value. Accept the governance centralisation risk and the early-holder concentration.

**FET/ASI** if you want the most liquid, most institutionally accessible DeAI position and are less concerned about actual decentralisation. Be aware that the alliance has fractured (Ocean's exit and an unresolved lawsuit) and the technology integration remains incomplete. The token will likely trade on narrative and partnerships more than on-chain fundamentals.

Of the three, MOR is the only zero-insider fair launch, and that alignment is the strongest case for the model even with the thin liquidity. TAO's sticking points are governance centralisation, extreme stake concentration and an emission-subsidised revenue model: independent analytics firm Pine Analytics ([24 March 2026](https://pineanalytics.substack.com)) puts identifiable external revenue across the entire Bittensor network at $3-15M, implying a revenue multiple one to two orders of magnitude above what centralised AI infrastructure peers command. FET's insider allocation and the messy ASI Alliance governance are the drawbacks there.

## Frequently asked questions

**How do Bittensor and Fetch compare?**
Bittensor (TAO) is a fair-launch mining network with a 21 million cap and no premine, but stake is heavily concentrated: on the median subnet a single coldkey holds a majority of it. Fetch (FET, now the ASI Alliance token) is migration-and-insider-weighted, governed through a Singapore company with a 91-validator cap. Bittensor is more decentralised by launch and mechanism; both are concentrated in practice. The [head-to-head](#bittensor-vs-fetch-head-to-head) above works through it.

**Which has the better token model, Bittensor or Fetch?**
Neither is a clean community launch. Bittensor's competitive-mining mechanism is more innovative and its hard cap gives a stronger scarcity narrative, but the stake concentration works against the decentralisation thesis. Fetch has broader reach and deeper liquidity, with insider-heavy economics and a merged tech stack that still runs as separate platforms. Of the three tokens here, only Morpheus launched with zero insider allocation.

**Is FET the same as ASI?**
Effectively yes. Fetch.ai merged into the Artificial Superintelligence (ASI) Alliance in 2024 with SingularityNET and Ocean Protocol. The token still trades as FET; the planned migration to an ASI ticker hasn't completed, and Ocean has since left the alliance.

*Holdings and disclosures are centralised on the [disclaimer](https://ownyourmind.ai/disclaimer/) page.*

## Score change log

| Date | Type | Score change | Notes |
| - | - | - | - |
| 2026-10-20 | Data | N/A | TAO's emission split restated as an 18/41/41 default that subnet owners can reset since eTAO (spec 481, 19 October 2026). |
| 2026-09-25 | Rating | Morpheus practical decentralisation High → Moderate | One 5-of-9 Safe holds owner and upgrade authority over the core contracts with no timelock, Governance scores 10/20 and no Snapshot venue remains. MOR still leads on Freedom; TAO's rating is unchanged. |
| 2026-09-25 | Data | N/A | Fetch validator cap corrected from 70 to 91; the set is small in practice because two validators can halt the chain. |