# FLock.io

> Decentralised AI training via federated learning on Base. NeurIPS-awarded, EF-granted. 11.9M FLOCK in protocol fees over 10 months from 16 training tasks. Academic rigour meets early-stage commercial traction.

- Scores: Freedom 5.5/10 (grade C), Returns 5.2/10 (grade D)
- Category: compute · Token: FLOCK
- Published: 2026-03-06 · Updated: 2026-10-07
- Canonical: https://ownyourmind.ai/projects/flock/
- Data (scores + integrity verdicts, JSON): https://ownyourmind.ai/api/projects/flock.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

FLock.io is a decentralised AI model training and validation platform that combines federated learning with blockchain. The core promise: train AI models collaboratively without sharing raw data. Only model updates (gradients) are transmitted; the training data never leaves the participant's device.

The architecture has three layers:

- **AI Arena** (train.flock.io): the live training platform on Base. Training Nodes submit models, Validators evaluate them via comparative scoring, and Delegators stake tokens to nodes and validators. On-chain smart contracts handle rewards and slashing.
- **FL Alliance**: the federated learning layer where participants fine-tune consensus models using their private local datasets. Still in testnet as of mid-2025. This is the privacy-preserving component that sets FLock apart.
- **Moonbase / API Platform**: deployment layer launched September 2025. An OpenAI SDK-compatible API for deploying and monetising trained models, with revenue sharing for contributors. Moonbase also introduced **FOMO (FLock Open Model Offering)** as part of Tokenomics v3: fair-launch model tokens backed by hosting services on FLock API. Model tokenholders get discounted usage or can stake for yield.

The technical innovation is **FLoRA**: using Low-Rank Adaptation (LoRA) for efficient fine-tuning with minimal compute overhead, replacing central aggregator servers with blockchain-based peer review. FLock achieved the first empirical validation of fine-tuning a 70B LLM in a decentralised multi-domain setting.

Founded by **Jiahao Sun** (MSc Computer Science, Oxford; Honorary Research Fellow, Imperial College London; former Director of AI at RBC Wealth Management). **Zehua Cheng** (Chief AI Scientist, DPhil candidate at Oxford) and **Dr. Zhipeng Wang** (Lead Blockchain Researcher, leading Ethereum Foundation-funded research) round out the technical leadership.

The entity is **FLOCK.IO LTD**, registered in London (Companies House #14039622, incorporated April 2022). Total funding: ~$9M, split between seed ($6M led by Lightspeed Faction and Tagus Capital, with DCG, OKX Ventures, Volt Capital) and strategic ($3M from Animoca Brands).

The academic credentials are strong: NeurIPS 2022 Best Paper Runner-up, NeurIPS 2023 Best Technical Demo, IEEE GBC 2025 Best Application Award, sole AI infrastructure recipient of an Ethereum Foundation Academic Grant (2024), and CB Insights AI 100 (2025).

## Value proposition

FLock addresses a hard problem. Most AI training requires centralising data, a privacy, sovereignty, and regulatory nightmare. Federated learning trains models where the data lives. FLock adds blockchain incentives (staking, slashing, rewards) to make this economically viable and verifiable.

### Usage signals: revenue, participation, but tiny task volume

The usage metrics FLock reported for its first 10 months (2025 Earnings Report, to 31 October 2025) tell a mixed story:

- **Protocol fee revenue: 11.9M FLOCK.** Revenue from Base smart contracts. FLock reported this in its 2025 Earnings Report (period to 31 October 2025) denominated in tokens, and that is how we carry it.
- **784,137 validation submissions** vs 9,062 training submissions, showing meaningful validator activity.
- **196 training nodes** including 6 top-50 Kagglers, a quality signal.
- **262 validators** and **1,416 delegators**, showing substantive participation.
- **55,928 token holders.**

But: only **16 training tasks completed** in that period. Per-task economics are impressive (roughly 740,000 FLOCK per task) but the volume is tiny. Task creation is still centralised: the FLock team decides what gets trained. Community task creation is planned but not shipped.

The gmFLOCK staking mechanism (Tokenomics v2, launched April 2025) is well-designed. Stake FLOCK to receive non-transferable gmFLOCK, required for AI Arena participation. Lock up to 365 days; longer lock means more gmFLOCK. FLock reports 25% of circulating supply voluntarily locked at an average of 265 days. That's demonstrable long-term alignment.

The healthcare use case, including a partnership with Moorfields Eye Hospital for ophthalmology diagnostics and cross-continental collaboration on blood glucose prediction for diabetic patients, demonstrates the real-world applicability of the approach.

### The counter-narrative: testnet status, closed-source, locked supply

The counter-narrative: FL Alliance, the core privacy-preserving federated learning layer, is still in testnet. The API Platform launched only in September 2025. Core platform code is partially closed-source. No smart contract audit has been published. And more than half of token supply is still locked, with significant future dilution ahead.

Partnerships with Qwen (Alibaba Cloud) and Animoca Brands are at the MoU stage, not revenue-generating agreements. More substantive is FLock's role as "AI Strategic Partner" in the UNDP SDG Blockchain Accelerator (second cohort, announced August 2025 alongside Stellar), mentoring five pilot projects across climate finance, energy access, healthcare, and supply chain. FLock participated in the UNDP High-Level Strategic Dialogue at UN Headquarters. That's a credibility signal that carries weight beyond crypto-native partnerships.

FLock also partnered with Walrus Protocol and Sui Foundation. Walrus provides decentralised encrypted storage for gradients, and FLock is working on fine-tuning an open-source model for the Sui network. This expands the chain footprint beyond Base.

## Tokenomics

FLOCK has a fixed supply of 1 billion tokens, a cap the deployed contract enforces, though a 2-of-3 Safe can upgrade that contract. The current float is shown in the fact strip above. TGE was December 31, 2024 on Base.

Distribution: community incentives 47%, team & shareholders 22.5%, treasury/ecosystem 18%, community airdrop 12.5%. The 66.7% community vs 33.3% insider split is better than most. Team and investor tokens have a 1-year cliff + 2-year linear vesting. Community tokens mint over 60 months with 1% monthly decay.

The gmFLOCK mechanism creates actual lock-up: stake FLOCK to receive non-transferable gmFLOCK (required for participation). Base lock of 30 days (1:1 ratio), scaling up with longer commitments (+0.006 per additional day). FLock reports 25% of circulating supply voluntarily locked at an average of 265 days.

No burn mechanism exists. Value accrual comes through protocol fee revenue (11.9M FLOCK in 10 months, per FLock) and planned revenue sharing via the API Platform. How fees flow to token holders vs the protocol treasury isn't explicitly documented.

FDV/MCap ratio signals meaningful future dilution; the current float is in the fact strip above. Listed on Coinbase, Gate.io, Bybit, KuCoin, MEXC, Bitget, and CoinEx. ATH was $0.667 in September 2025 (the token now trades around 90% below it).

## How to participate

**Run a Training Node.** Submit AI models to training tasks. Stake FLOCK via gmFLOCK. Includes top-50 Kagglers among current participants. ML expertise required. Technical skill: advanced.

**Stake as Validator or Delegator.** Validators evaluate training submissions. Delegators stake to trusted nodes and validators. Lock FLOCK via gmFLOCK for up to 365 days. Technical skill: intermediate.

**Deploy Models.** Use the API Platform (Moonbase) to deploy trained models. Revenue sharing for contributors. Elite Trainer Programme offers 50% revenue share. Technical skill: intermediate.

## Honest assessment

## Freedom Score: 55/100

FLock has a federated learning architecture inherently designed for data privacy, strong academic credentials, and decent community token distribution. But task creation is centralised, core code is partially closed, and governance is aspirational.

**Infrastructure Decentralisation: 11/20.** FLock reported 196 training nodes and 262 validators to October 2025, providing verifiable multi-party training. Task creation remains centralised to the FLock team. FL Alliance (the privacy-preserving federated learning layer) still in testnet. Smart contracts on Base inherit Ethereum security.

**Governance Decentralisation: 9/20.** FLock's docs describe DAO governance as an initial vision, subject to change before launch. In practice, team controls task creation, protocol upgrades, and treasury. A decentralisation plan was due Q2 2025, but no evidence it was published. Governance is more described than demonstrated.

**Token Distribution Fairness: 10/15.** 66.7% community vs 33.3% team/investors, a good ratio. 1-year cliff + 2-year linear vest for team. Community tokens minted over 60 months with decay. 5% airdrop to testnet participants. 48.7% circulating after about 21 months is a moderate unlock pace. (figures as of 2026-10-19)

**Censorship Resistance: 8/15.** Federated learning means data never leaves participants' devices, a strong privacy architecture by design. The token contract is a different story. It sits behind a L2 proxy that a 2-of-3 multisig can replace with no timelock, and it carries an admin-controlled blacklist that blocks transfers from any address the admins name. Task creation is centralised and the API Platform is team-operated too, so there are single points of control over what gets trained, what gets served, and who can move their tokens.

**Data Sovereignty: 10/15.** Core value proposition is data staying with owners. FL Alliance designed for privacy-preserving training without data movement. Healthcare demo validated cross-continental collaboration without data sharing. FL Alliance still in testnet, so the full vision isn't yet delivered.

**Open Source Transparency: 7/15.** 33 non-fork GitHub repos, 12 of them with no licence file, and some actively maintained. Training node quickstart and validator code are open source. Main FLock repo is near-empty (README + LICENSE only). Core platform code and smart contracts not fully public. Academic papers peer-reviewed at NeurIPS and IEEE add credibility.

## Returns Score: 52/100

**Token Utility: 14/20.** FLOCK required for staking as trainer, validator, or delegator. gmFLOCK mechanism adds meaningful lock-up incentive (up to 365 days, with FLock reporting 25% of circulating supply voluntarily locked). Governance voting requires staking. Slashing for malicious behaviour creates concrete skin-in-the-game. Task creation will require tokens (planned). No burn mechanism.

**Value Accrual: 10/20.** The case rests on FLock's self-reported figure of 11.9M FLOCK in fee revenue over its first 10 months, from its 2025 Earnings Report. No burn or buyback exists, and FLock leaves undocumented whether fees reach holders or stay in the treasury. The fee flow itself has stalled, as the next section shows. API Platform revenue sharing and the Elite Trainer Programme's 50% share reward contributors, and neither routes value to passive holders.

**Supply Dynamics: 14/20.** 1B fixed supply with no additional inflation. Community tokens minted over 60 months with 1% monthly decay, a controlled emission schedule. FLock reports 25% of circulating voluntarily locked via gmFLOCK (average 265 days). Team/investor cliff-locked for 1 year + 2-year linear vest. However, more than half of supply is still locked, creating significant future dilution pressure.

**Revenue Sustainability: 5/25.** Fees have effectively stopped. DefiLlama's last 30 days of FLock revenue annualise to under 1% of its trailing-year figure on a 24 September 2026 read.  The self-reported 11.9M FLOCK came from only 16 training tasks, and task creation is still centralised, with community task creation not yet live.

The partnerships haven't changed that. UNDP is active (AI Strategic Partner in SDG Blockchain Accelerator), while Animoca and Qwen remain at MoU stage. The API Platform, launched in September 2025, has no separate revenue figure in this review.

Two things qualify the headline. DefiLlama's adapter reports about $544,000 in fees over the trailing year and $291 over the trailing 30 days on a 21 September 2026 read, so fee generation has effectively stopped rather than slowed. And the token-denominated total is the only honest way to carry it: FLock reported fees in FLOCK, and the dollar conversion depends entirely on which day you pick. Both are flagged for the next monthly review.

**Liquidity & Access: 9/15.** Listed on Gate.io, Bybit, KuCoin, MEXC, CoinEx, Bitget and Coinbase, across dozens of CoinGecko tickers, providing decent CEX coverage for a small-cap project. DEX liquidity on Uniswap Base and Aerodrome. Volume is the opposite of thin and that is its own warning: the 7-day median turnover ran at 243% of circulating market cap in early September and about 98% on 24 September, and about half that flow is Upbit,  so this is Korean retail rotation rather than depth.

No Binance listing. OKX added a FLOCK/USDT perpetual at up to 20x on 12 September 2026, which brings margin trading but no spot book. Base chain primary, with Walrus/Sui integration expanding the chain footprint.

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

FLock sits in Quadrant B: substantive data sovereignty innovation with early commercial traction, but too small and too early to demonstrate sustainable returns.

### Key risks

- **16 training tasks in 10 months.** Fees per task are strong (roughly 740,000 FLOCK) but the volume is tiny. Can it scale 100x?
- **Centralised task creation.** The team decides what gets trained. Community task creation is planned but not shipped.
- **More than half of supply locked.** Significant future dilution as team, investor, and community tokens unlock over the next 3+ years.
- **FL Alliance still in testnet.** The core privacy-preserving federated learning layer (the key differentiator) isn't yet in production.
- **Core platform partially closed-source.** Smart contracts and platform logic not fully public despite open source claims.
- **No smart contract audit.** A protocol handling staking and slashing with no published security audit is a concern.
- **Small market cap.** Vulnerable to large sells. Volume means limited exit liquidity for meaningful positions.
- **Competition.** Bittensor, Gensyn, and others compete for decentralised AI training. FLock's federated learning approach is differentiated but unproven at scale.
- **Partnership execution.** UNDP is active, but Qwen and Animoca partnerships remain at MoU stage, not revenue-generating.

The core question for FLock is whether 16 training tasks in 10 months becomes 160 in the next 10. If FL Alliance ships to mainnet and community task creation launches, the per-task economics suggest meaningful upside. If not, the 11.9M FLOCK in protocol fees is a one-off data point from a still-centralised system.

## Score change log

| Date | Score | Change | Reason |
| - | - | - | - |
| 2026-10-07 | Data | N/A | Liquidity wording: 7-day median turnover fell from about 100% to about 32% of market cap in early October, so the frontmatter now gives the range. Scores unchanged. |
| 2026-10-04 | Data | N/A | Liquidity evidence adds OKX's FLOCK/USDT perpetual (up to 20x), listed 12 September 2026 per OKX's announcement. Scores unchanged. |
| 2026-10-04 | Returns | Liquidity & Access 8 → 9 (Returns 51 → 52, D held) | Coinbase lists FLOCK and turnover runs near a full circulating cap a day, though about half of it trades on Upbit. |
| 2026-09-25 | Score | Returns 57 → 51 (C → D) | Revenue Sustainability 8 → 5: DefiLlama's last 30 days annualise to under 1% of the trailing year. Value Accrual 13 → 10: it rested on the self-reported 11.9M FLOCK, with no burn or buyback and an undocumented fee destination. |
| 2026-09-24 | Data | N/A | Phase 5 sweep: the 11.9M FLOCK fee and usage figures attributed to FLock's 2025 Earnings Report (no longer "verified on-chain"), trailing fees refreshed to DefiLlama's September read, locked share, turnover, repo count and venue wording refreshed, past unlock date dropped. Scores unchanged. |
| 2026-09-23 | Data | N/A | Integrity records re-cut: 2 claims now quote FLock's own pages verbatim and 5 are our own measurements. Scores unchanged. |
| 2026-09-07 | Data | N/A | WS3b open-source rail: our evidence checks out, including the near-empty main repo. Adds that 12 of 33 repos carry no licence and 25 have gone six months untouched, while the validator the network depends on was pushed this week. Open Source 7/15 held. |
| 2026-09-06 | Data | N/A | Research JSON resynced to the frontmatter Returns breakdown. Revenue Sustainability read 9 in the JSON after the 2026-09-05 move to 8 never propagated. No published score changed. |
| 2026-09-06 | Data | N/A | WS3b liquidity rail: two corrections. Coinbase does list FLOCK (FLOCK-USD online, checked at Coinbase), against our no-Binance-or-Coinbase line, and turnover is a 242.98% 7-day median with Upbit at 49.7% of flow, against our volume-adequate-but-not-deep line. Liquidity & Access 8/15 flagged for October. |
| 2026-09-05 | Returns | Revenue Sustainability 9 → 8 (Returns 58 → 57, C held) | Trailing-year revenue fell 26.3% between 11 Jun and 16 Aug 2026 once the basis was rebased off a cumulative all-time total, and the last 30 days annualise to 7% of the trailing year. The token-denominated total was never wrong; the dollar headline was a frozen price conversion. |
| 2026-08-16 | Data | N/A | Corrected "Base L2 contracts are immutable": the FLOCK token is a transparent proxy upgradeable by a 2-of-3 Safe with no timelock, and it carries an admin-gated blacklist that blocks transfers. Censorship Resistance 8/15 and Governance 9/20 flagged for monthly review. |
| 2026-08-15 | Data | N/A | Circulating supply de-hardcoded and bound. The stated ~41% was 2.8pp stale against 43.75% of max supply. |
| 2026-08-10 | Data | N/A | WS3 holder-source: authored token_distribution_fairness (`match`, onchain). Base top-10 ex-infra ~44% = D1 moderate, consistent with the claimed 66.7% community / 33.3% team ratio (10/15). Re-check flagged: the largest holder is a 28.95% unlabelled EOA that could shift the read if it is a team wallet. No score change. |
| 2026-08-17 | Data | N/A | Correction. The $2.7M revenue headline was 11.9M FLOCK converted at an Oct-2025 price; the same tokens are worth ~$355K today. All 14 surfaces re-denominated in FLOCK, per the price-agnostic rule. Scores flagged for the monthly. |
| 2026-06-26 | Data | N/A | DefiLlama LlamaAI harvest cross-check, verified independently (CoinGecko + api.llama.fi). Refreshed circulating to ~41% (~410M of 1B, from 38.7%) and the locked share to ~59% (from 68.5%) on continued vesting; mainnet age to ~18 months; added the ATH figure ($0.667, Sep 2025, ~95% below). Harvest surfaced a revenue-deceleration signal worth recording: DefiLlama's trailing-12-month tracked fees are ~$1.07M and the on-chain run-rate is down ~52% month-on-month, so the $2.7M cumulative FLOCK-denominated headline overstates the current pace. Revenue Sustainability (9/25) held this pass but flagged for monthly-review reassessment. No score change. |
| 2026-06-02 | Data | N/A | Circulating supply corrected from 31.5% (~315M) to 38.7% (~387M of 1B) per CoinGecko; "14 months" → "17 months" since TGE; removed stale "3.16x" FDV/MCap absolute. Score unchanged. June monthly review. |
| 2026-04-06 | Data | N/A | Added UNDP SDG Blockchain Accelerator partnership, Walrus/Sui integration, FOMO (Tokenomics v3), Moorfields Eye Hospital naming. No score changes. |
| 2026-03-24 | Data | N/A | Editorial data review and correction. Verified against on-chain sources. |
| 2025-03-06 | Both | N/A | Initial publish. Freedom 60/100, Returns 58/100. |