# Allora Network

> Self-improving decentralised AI inference network using regret-based model aggregation on Cosmos SDK. Tier-1 exchange listings, $35M raised, zero verified revenue, massive unlock cliff November 2026.

- Scores: Freedom 4.7/10 (grade D), Returns 4.6/10 (grade D)
- Category: other · Token: ALLO
- Published: 2026-03-06 · Updated: 2026-10-04
- Canonical: https://ownyourmind.ai/projects/allora/
- Data (scores + integrity verdicts, JSON): https://ownyourmind.ai/api/projects/allora.json
- Cite as: Own Your Mind (ownyourmind.ai), licence CC BY 4.0. Link the canonical URL.

Charts and interactive widgets are omitted here; their data is in the JSON above.

---

## What it does

Allora Network is a self-improving decentralised AI inference and prediction network. It aggregates competing machine learning models to produce collective intelligence that outperforms any individual model, like a prediction market for AI outputs.

Built on Cosmos SDK with CometBFT consensus, the network is organised around **Topics**: compartmentalised sub-networks targeting specific prediction problems (BTC price prediction, yield optimisation, DeFi risk assessment). Each Topic has four participant types:

- **Workers** generate ML inferences by deploying models and submitting predictions
- **Forecast Workers** predict the accuracy of other workers' outputs
- **Reputers** evaluate all outputs against ground truth data and score quality
- **Validators** secure the underlying chain

The novel mechanism is **Inference Synthesis**: a multi-step aggregation process using regret-based weighting. Inferences receive weights based on whether they outperform previously calculated predictions. This prevents initially successful predictors from gaining disproportionate influence and allows the system to continuously improve as better models join.

Getting onto mainnet as a worker is gated. Allora's 13 August 2026 Allora Forge post describes promotion and relegation, run per worker and per topic against each topic's testnet twin. Promotion needs a pass on seven tests covering sample size, directional accuracy above 50%, correlation, beating a flat baseline on two loss metrics, forecast scaling and participation above 90%. Relegation needs strong joint evidence of failure at 95% confidence, or prolonged inactivity, after a 30-day grace period.

The chain enforces the gate. Our 24 September 2026 read of the mainnet emissions module found worker and reputer whitelists on for all 23 topics and topic creation limited to whitelisted accounts, and allora-chain's `CanSubmitWorkerPayload` rule admits only listed addresses to a whitelisted topic.  So a promoted worker serves only once the topic creator or the chain's whitelist admins list its address, and the post calls the deciding system "backend logic" without saying who runs it. A fixed set of seven thresholds is also a target a worker can tune a model to clear.

The same gate applies to outside integrations. [peaq](https://ownyourmind.ai/projects/peaq/)'s robotic.sh adapter, announced on 3 September 2026, lets a machine pay the 1 ALLO mainnet fee and register as an Allora worker, but registration alone does not clear the whitelist, so a registered machine cannot submit to a mainnet topic until it is listed.

**Pay-What-You-Want (PWYW)** is the fee model. Consumers determine what they pay for inferences. Topics with zero payment receive zero emissions weight, effectively deprioritising them. Consumer contracts are live on Arbitrum One and Sepolia, enabling cross-chain consumption.

The project started as **Upshot** in 2019, focused on NFT appraisal using peer prediction. Founded by **Nick Emmons** (CEO) and **Kenny Peluso** (CTO, BSc Applied Mathematics, Brown University). They recognised broader applications for their peer prediction technology and rebranded to Allora Labs in February 2024. The entity is Allora Labs Inc (formerly Upshot Technologies), registered in the Cayman Islands with operations in New York and Boston. Team of approximately 30 people.

Allora Labs reports $35 million raised across three rounds: Series A ($7.5M, 2021), Series A Extension ($22M, 2022), and Strategic ($3M, Archetype-led, 2024). Investors listed include Polychain Capital, Framework Ventures, Blockchain Capital, CoinFund, Delphi Digital, and notable angels Stani Kulechov (Aave), Kain Warwick (Synthetix), and Ryan Selkis (Messari).

Mainnet launched 11 November 2025 alongside TGE, with simultaneous Binance, Coinbase, and Kraken listings, an exceptionally rare achievement for a DeAI project. The token crashed over 50% on day one.

## Value proposition

Decentralised AI inference that gets smarter over time. The core thesis is that aggregating multiple competing ML models produces better predictions than any single model.

The regret-based Inference Synthesis mechanism is novel. Unlike Bittensor's subnet model where miners compete for validator approval, Allora's system mathematically weights inferences based on their marginal contribution to prediction accuracy. Models that consistently add unique value are weighted higher. The system is designed to resist Sybil attacks: running the same model twice doesn't increase rewards because the second copy adds no unique information.

The DeFi integrations are live and growing. PancakeSwap uses Allora predictions for price prediction games. Steer Finance powers AI-driven vaults. Drift Protocol runs perps agent vaults. Coinbase has integrated Allora as a CDP AgentKit action provider. Alibaba Cloud is a strategic partner. Cross-chain consumer contracts bring Allora predictions to Ethereum, Solana, Arbitrum, Base, Monad, and TRON.

Self-reported metrics: 692 million inferences generated, 288,000 workers, 55 active topics. However, these figures include testnet activity and are likely inflated by airdrop farming during the testnet phase. Mainnet-only metrics are not publicly broken out.

### Two risks that dominate

Two risks matter more than everything else combined.

**First: the PWYW model.** Consumers set their own prices for inferences. If the market converges on near-zero fees, the token economy collapses regardless of prediction quality. Topics with zero payment receive zero emissions weight, which theoretically self-corrects, but it also means the network could have many active topics generating inferences with minimal revenue. There's no independently verified revenue data to assess whether paying demand exists.

**Second: the vesting cliff.** 48.55% of total supply (Early Backers 31.05% + Core Contributors 17.50%) begins unlocking in November 2026. The first 33% tranche is worth roughly two-thirds of the current circulating supply, so it could lift the float by about that much in a single event. Which supply the 48.55% is measured against matters: not all of the 1 billion max has been issued, so the same unlock reads as about a quarter of max supply or nearer a third of what actually exists.  The token is down more than 70% from its ATH, having recovered off its day-one lows.

For the sovereignty thesis, Allora's architecture is designed for data sovereignty: workers retain ownership of their ML models without uploading weights. Claims of zkML integration for verifiable inference without exposing proprietary models are promising but implementation status is unclear. The Cosmos SDK foundation provides decentralisation potential, though a validator set in the teens is highly concentrated, and the threshold that matters is lower again: 5 of them hold more than a third of stake between them. (figures as of 2026-10-19)

## Tokenomics

ALLO has a maximum supply of 1 billion tokens with 255.1 million circulating (25.5% of max). The emission model is disinflationary, with Bitcoin-like decreasing emissions over time. (figures as of 2026-10-19)

Distribution is heavily insider-weighted:

- **Early Backers: 31.05%**, 12-month cliff (from November 2025), then 33% unlock, remainder linear over 24 months. Currently 0% unlocked.
- **Network Emissions: 21.45%**, decreasing emissions over time.
- **Core Contributors: 17.50%**, same cliff and vesting as Early Backers. Currently 0% unlocked.
- **Foundation: 9.35%**, ~50% at TGE, remainder 24-month linear. 56.6% unlocked.
- **Community Pool: 9.30%**, 100% at TGE. Distributed for qualifying testnet activities.
- **Ecosystem & Partnerships: 8.85%**, 50% at TGE, 50% 24-month linear. 54.2% unlocked.
- **Allora Prime Staking: 2.50%**, 100% at TGE. Enhanced 9-month staking programme.

The critical event: **November 2026**, when Early Backers (31.05%) + Core Contributors (17.50%) = 48.55% of total supply begins unlocking. The first 33% tranche of approximately 160 million tokens could lift the current circulating supply of 255.1 million by roughly two-thirds. Series A investors entered at $60M pre-valuation and the extension at $220M pre-valuation. Compared to current FDV, extension investors are significantly underwater. (figures as of 2026-10-19)

No burn mechanism documented. Inference fees flow before new tokens are minted, reducing effective inflation as usage rises. Excess fees accumulate in an Emissions Treasury.

ALLO's volume-to-market-cap ratio sits mid-pack against our own corpus, with Binance the largest venue. Listed on Binance, Coinbase, Kraken, OKX, KuCoin, Gate.io, and MEXC. Only about a quarter of max supply is circulating, and an FDV/MCap ratio above 3x reflects significant dilution ahead.

Staking offers approximately 12% APY base, with Allora Prime providing up to 50% APY (12% base + 38% Prime rewards) for eligible participants during a 9-month programme. 21-day unbonding period. Staking rewards are emission-funded, not revenue-funded.

## How to participate

**Stake ALLO.** Delegate to validators or reputers via the Allora Explorer. Earn ~12% APY base or up to 50% with Allora Prime (time-limited). 21-day unbonding period. Technical skill: basic.

**Run a Worker Node.** Deploy ML models to generate inferences for network Topics. Earn rewards based on inference accuracy and unique contribution. Must stake ALLO as commitment bond, and a mainnet slot comes only after passing the testnet promotion tests described above. Technical skill: expert (requires ML expertise).

**Operate as a Reputer.** Evaluate inference quality against ground truth data. Requires access to reliable ground truth data sources. Must stake ALLO. Technical skill: advanced.

**Consume Predictions.** Integrate Allora predictions into DeFi applications or agents via SDKs (Go, Python, TypeScript) or consumer contracts on Arbitrum. PWYW fee model. Technical skill: intermediate.

## Honest assessment

## Freedom Score: 47/100

16 active validators and 48.55% of tokens in insider hands. Cosmos SDK gives Allora a sound foundation; what's built on it right now is not particularly decentralised. (figures as of 2026-10-19)

**Infrastructure Decentralisation: 7/20.** Cosmos SDK L1 on CometBFT, but only 16 active validators, 5 of which can halt the chain. Workers are whitelisted rather than permissionless: our 24 September 2026 read found worker and reputer whitelists on for all 23 mainnet topics, and only listed addresses can submit. Off-chain ML execution without deployed zkML adds further centralisation vectors.  (figures as of 2026-10-19)

**Governance Decentralisation: 7/20.** ALLO holders can vote on protocol upgrades, emission parameters, and community grants, and the chain shows they do. Seventeen proposals have gone through the governance module: validator-set changes, fee-market settings, an IBC light-client recovery, the vote that switched emissions on, and six protocol upgrades through v0.17.0 on 13 August 2026. Proposal 13 was voted down and the same upgrade then passed as proposal 14, so the process can say no.

Participation is the weak part. Every proposal that passed drew zero opposing votes bar that one rejection, and 48.55% of supply is locked to insiders, so the Foundation, the Community Pool and the largest insider allocations dominate any tally. The process works and is barely contested.

**Token Distribution Fairness: 5/15.** Early Backers (31.05%) + Core Contributors (17.50%) = 48.55% insider allocation. Community Pool is only 9.3%. Binance airdrop was 1.5% of supply. Heavy insider concentration despite 3-year vesting schedules.

**Censorship Resistance: 6/15.** Participation is gated on chain. Creating a topic needs a whitelisted account, and every mainnet topic enforces worker and reputer whitelists, so whoever controls the lists controls who can take part. With 16 validators and 5 able to halt the chain, coordinated censorship is feasible, and the PWYW model lets topics be economically starved. (figures as of 2026-10-19)

**Data Sovereignty: 10/15.** Workers retain ownership of ML models, with no uploading of weights required. zkML claims to verify model execution without exposing proprietary weights, though implementation status unclear. Architecture is designed for data sovereignty by default.

**Open Source Transparency: 12/15.** The core chain is open source under Apache 2.0 licence. The org carries 11 non-fork public repositories (14 including forks), 6 of them with no licence file, with allora-chain the core node. SDKs in Go, Python, TypeScript. Whitepaper publicly available. A Halborn audit was announced in July 2024 and no completed report has been published by either party, which is a significant transparency gap for a network handling financial predictions.

## Returns Score: 46/100

**Token Utility: 14/20.** ALLO is structurally required for inference purchases, topic creation, worker/reputer registration bonding, staking, and governance. Multi-dimensional utility. However, the PWYW model means consumers set their own prices. If the market converges on low fees, token demand for inference could be minimal. Staking creates lock-up demand. Score reflects structural utility tempered by PWYW demand uncertainty.

**Value Accrual: 10/20.** Inference fees flow to workers and reputers, not directly to passive token holders. No fee burn or buyback mechanism documented. Value accrual is primarily through staking rewards (~12% APY) which are emission-funded, not revenue-funded: the rewards inflate supply rather than distributing fees. Allora Prime offers up to 50% APY but is time-limited and also emission-funded. The protocol needs significant inference fee volume to create value accrual beyond dilution. No independent evidence of meaningful fee revenue.

**Supply Dynamics: 8/20.** Critical overhang: 48.55% of supply begins cliff unlock in November 2026. The 33% initial unlock represents ~160M tokens versus 255.1 million current circulating, potentially lifting float by roughly two-thirds. An FDV/MCap ratio above 3x indicates significant dilution ahead. Only about a quarter of max supply circulating. Disinflationary emissions are positive long-term but near-term unlock pressure is severe. (figures as of 2026-10-19)

**Revenue Sustainability: 3/25.** No independently verified protocol revenue, and neither DeFiLlama nor Token Terminal tracks it. The PWYW model means consumers can pay zero for inferences. Topics with zero payment receive zero emissions weight, but this creates a race-to-bottom risk. Network metrics (692M inferences, 55 topics) are self-reported and include testnet. Revenue from paying consumers on mainnet is not documented. Revenue sustainability is the biggest question mark.

**Liquidity & Access: 11/15.** Listed on all major exchanges simultaneously at launch: Binance, Coinbase, Kraken, OKX, KuCoin, Gate.io, MEXC, with Binance holding a little over a quarter of the flow and no second venue above a fifth. That coverage is genuinely strong for a token this size. Turnover, measured the same way on the same day across every project we track, sits mid-pack rather than near the top: Virtuals and Aethir both read higher, before counting the tokens whose ratios are one-venue artefacts.

### Quadrant: B (High Freedom, Low Returns)

Allora is a bet on the PWYW model finding equilibrium before the November 2026 cliff arrives. If paying customers materialise and inference fee volume builds, the disinflationary emission structure starts doing useful work. If the market converges on near-zero fees before that, the cliff event gets ugly. The regret-based mechanism is interesting; the economics haven't proved out yet.

### Key risks

- **Massive vesting cliff.** 48.55% of supply unlocks starting November 2026. First 33% tranche (~160M tokens) could lift circulating supply by roughly two-thirds. This is the defining risk event.
- **Zero verified revenue.** PWYW model allows zero-fee consumption. Not on DeFiLlama. No independent revenue data. Economic sustainability entirely unproven.
- **A validator set in the teens.** 16 active validators is highly concentrated for an L1 chain, and 5 together can halt it. (figures as of 2026-10-19)
- **Over 70% from ATH.** Debuted high, crashed 50%+ day one, since recovered off the lows. All investors from the $220M valuation round deeply underwater.
- **No published security audit.** Allora commissioned one: it announced a Halborn engagement in July 2024, scheduled to complete before mainnet launch. Mainnet launched and no report has appeared, on Allora's side or in Halborn's own public audit index, which we walked in full on 18 September 2026. A commissioned audit with no published report is a different thing from never having sought one, and neither gives a reader anything to check.
- **PWYW demand uncertainty.** If consumers converge on near-zero fees, the token economy collapses regardless of prediction quality.
- **Pivot history.** Upshot (NFT appraisal) did not achieve product-market fit. The pivot to AI inference is legitimate but the pattern is concerning.
- **Self-reported metrics.** 692M inferences and 288K workers include testnet and airdrop farmers. Mainnet-only activity unclear.

## Score change log

| Date | Score | Change | Reason |
| - | - | - | - |
| 2026-10-04 | Returns | Liquidity & Access 12 → 11 (Returns 47 → 46, D held) | Turnover is mid-pack in our corpus, not among the highest in DeAI as the evidence once said. |
| 2026-09-24 | Score | Freedom 52 to 47 | Infrastructure 9 to 7 and Censorship 9 to 6: the rationales credited permissionless workers and open topic creation, but mainnet enforces worker, reputer and topic-creator whitelists on all 23 topics. |
| 2026-09-24 | Data | N/A | Validator count and halt coefficient bound live (16 active; the halt threshold is now five, not six); circulating supply bound live (250M, up from the hand-typed 235M); FDV/MCap and ATH drawdown restated structurally; removed the stale "eight months away" cliff dating. Scores unchanged. |
| 2026-09-24 | Data | N/A | Added Allora Forge (13 Aug 2026): workers reach mainnet by passing seven testnet tests. Our chain read shows worker whitelists on for all 23 mainnet topics and topic creation whitelisted, which conflicts with the "permissionless" wording; flagged for the monthly. Noted peaq's worker-registration adapter. Scores unchanged. |
| 2026-09-23 | Data | N/A | Integrity records re-cut: 3 claims now quote Allora's own pages verbatim and 6 are our own measurements. Scores unchanged. |
| 2026-09-18 | Data | N/A | `audit_status` moved unverified → outdated: Halborn's full public audit index (12 pages, 347 entries, control-checked) carries no Allora entry, and the announced report remains unpublished two years on. Scores flagged for the monthly review. |
| 2026-09-07 | Data | N/A | WS3b consensus rail: our "16 validators could coordinate censorship" line described the set size, not the threshold. Six colluding validators exceed a third of stake, and barely more than one percent of supply is bonded. Corrected on three surfaces. Censorship 9/15 flagged for October. |
| 2026-09-07 | Data | N/A | WS3b supply rail: three surfaces carried three different FDV/MCap ratios and three different float figures, all drifted in the project's favour, and the body dated the November cliff as "eight months from now". Only 788M of the 1B max is issued, so the denominator changes the headline. The tranche-to-float ratio held. Supply Dynamics 8/20 held. |
| 2026-09-06 | Data | N/A | WS3b liquidity rail: the volume-to-market-cap-among-the-highest-in-DeAI claim fails against our own corpus, where Virtuals and Aethir both read higher on the same measure the same day, so it is restated as mid-pack. All seven named venues confirm. Liquidity & Access 12/15 flagged for October. |
| 2026-08-16 | Data | N/A | Correction. The Governance evidence claimed no active proposals; the chain carries 17, including a rejection (prop 13) followed by the same upgrade passing (prop 14). On-chain verdict added; score flagged for the monthly. |
| 2026-08-12 | Data | N/A | WS3 open-source verdict added (1->2). allora-network/allora-chain is Apache-2.0, 140 stars, active (match); complements the existing native uallo supply reconciliation, which the new WS3 Cosmos source now automates. Scores unchanged. |
| 2026-06-26 | Data | N/A | DefiLlama LlamaAI harvest cross-check, verified independently (CoinGecko + api.llama.fi). The token has recovered since publish: ATH decline now ~76% (CoinGecko -76.3% from $1.60, 11 Nov 2025) versus the ~93% cited in March. Refreshed circulating supply to ~235M (~24% of the 1B max, up from ~220M/22% on emissions) and FDV/MCap to ~3.3x (from 3.65x, CoinGecko minted-supply basis). The Nov-2026 cliff arithmetic re-framed: the ~160M first tranche now lifts the float by roughly two-thirds rather than "nearly doubling" it. Revenue still zero/untracked on DeFiLlama; no score change. |
| 2025-03-06 | Both | N/A | Initial publish. Freedom 50/100, Returns 47/100. |