Q2 2025 marked a period of significant progress and validation for FLock.io. As the global conversation around AI continues to accelerate, FLock has taken concrete steps to demonstrate that decentralised, privacy-first AI is not only viable, but increasingly essential.
This quarter saw the launch of key technical initiatives, the formation of high-impact strategic partnerships, and growing recognition from both industry and policymakers. From our inclusion in the CB Insights AI 100 to our presence at leading global forums and events, FLock continues to establish itself as a foundational layer for the future of federated intelligence. The following update outlines our key milestones across product development, ecosystem engagement, and strategic expansion.
Milestones & Network Expansion
- FLock.io was named to the 2025 CB Insights AI 100, joining a globally recognised list of the most promising AI start-ups. As the only project on the list focused exclusively on open federated learning and collaborative model training, this honour underscores FLock’s leadership in building privacy-first AI infrastructure. It validates our unique approach in a field dominated by centralised incumbents and signals growing institutional confidence in decentralised AI.
- We officially launched FLock OFF, a permissionless federated learning subnet on Bittensor (UID 96), optimised for training Small Language Models (SLMs) on edge devices. This is a major step toward realising our vision of scalable, efficient, and democratized AI training, where anyone can contribute, validate, and benefit from decentralised intelligence. FLock OFF brings us closer to developing a lightweight open-source SLM that can outperform proprietary nano-models, giving the open AI movement real competitive footing.
- Our alliance with Alibaba Cloud’s Qwen represents a transformative collaboration between a global AI leader and a decentralized AI pioneer. Qwen brings world-class LLM performance and research depth, while FLock contributes privacy-first infrastructure and federated scalability. This partnership proves that decentralised AI isn’t just philosophically compelling — it’s now technically and commercially viable at the highest levels of AI development, opening doors for global enterprise adoption.
- The rollout of our Elite Trainer Programme marked a pivotal moment in community engagement. By onboarding world-class ML engineers into a dedicated development track, FLock is cultivating a high-performance contributor base. This not only accelerates technical innovation but also strengthens the quality, diversity, and resilience of our federated network — essential for building models that are open, robust, and community-aligned.
- This quarter also saw the introduction of gmFLock, a key advancement in FLock’s network design and tokenomics. gmFLock is a non-transferable token generated by staking FLOCK, serving as proof of participation across three critical ecosystem roles: Training Node Operator, Network Validator, and Delegated Stake Participant. The longer users stake FLOCK (up to 365 days), the more gmFLOCK they receive, increasing their influence and rewards from the network’s incentive pools. This upgrade transforms FLOCK into productive capital, encouraging sustained participation while reducing speculative volatility. By aligning economic incentives with long-term contribution, gmFLOCK strengthens the decentralisation, resilience, and sustainability of the entire FLock ecosystem.
- FLock.io launched decentralised exchange (DEX) liquidity for the $FLOCK token on Uniswap’s Base network, with newly established trading pairs: FLOCK/USDT, FLOCK/USDC, and FLOCK/WETH. This marks a significant step forward in strengthening the utility and accessibility of the FLOCK token across the ecosystem. By enabling greater flexibility for participants to engage in AI Arena tasks and other platform activities, this liquidity deployment supports a more open, user-driven economy. It also reinforces FLock’s commitment to building the first decentralised AI training network with meaningful, real-world token utility, further aligning economic infrastructure with our long-term mission to decentralise AI innovation.
- FLock.io won the Best Application Award at the IEEE Global Blockchain Conference 2025 for its ground-breaking work on blockchain-enabled federated learning in healthcare. Our award-winning paper demonstrated how hospitals across Europe, North America, and Asia could collaboratively train AI models—such as for blood glucose prediction—without sharing sensitive patient data. By integrating blockchain, we ensured transparency, fairness, and trust in the model training process. The award recognises FLock.io’s solution as not only innovative but also highly practical, with significant implications for privacy-preserving AI in global healthcare.
Strategic Partnerships & Ecosystem Growth
- As an extension of our strategic alliance with Alibaba Cloud’s Qwen, Base ecosystem projects get up to 50% off Qwen tokens using FLock-trained models or other major Qwen variants, plus: 1 free FLock training task and 1hr free AI consultation. This offering exemplifies our commitment to support AI development on Base. Projects can find the application form on the Base Services Hub.
- Our integration with SpoonOS, ****the agentic AI operating system for Web3, allows FLock-trained models to operate within intelligent, interoperable ecosystems. By linking through the Model Context Protocol and Agent-to-Agent Protocol, we enable seamless onchain coordination between AI agents. This expands our technical footprint across the agent economy and places FLock at the core of emergent, autonomous AI infrastructure — precisely where decentralised AI will thrive.
- The launch of Pay-fAI with Infini brings decentralised AI into the financial services arena, one of the most sensitive and privacy-critical domains. By powering InfiniGPT with FLock’s federated learning, we’re demonstrating that high-performance, secure AI is possible without centralising sensitive user or financial data. This project underscores FLock’s potential in regulated markets, where privacy, compliance, and transparency are non-negotiable.
- Our vehicle fuel efficiency task, developed in partnership with DIMO and Beacon Protocol, marks the first-ever AI model trained privately on community-owned driving data and will deliver personalized route optimization and fuel efficiency in real time to users.
Research & Development Initiatives
- We unveiled Baby4D, a first-of-its-kind prenatal tool that lets expectant parents preview their baby’s future face using AI and 4D ultrasound — all while preserving medical privacy. Baby4D is powered by FLock’s federated learning infrastructure, which keeps data local and secure. This showcases a real-world, deeply personal application of FLock’s core technology and opens up new verticals in privacy-first health-tech — a major growth frontier for decentralised AI.
- Our presentation at The Web Conference 2025 marked a historic turning point. As the first community-driven AI training platform on blockchain to be featured on the same stage that once launched Google’s earliest innovations, this event signalled mainstream academic and industry recognition of decentralised AI. For FLock, it was more than visibility — it was validation of our architecture, governance model, and technical depth in front of the most respected voices in AI research.
- FLock contributed to deep-dive workshops at ZuBerlinCity alongside Web3 visionaries like Vitalik Buterin and Martin Köppelmann. These sessions explored the future of decentralised coordination, trust, and DAO-powered innovation. Our participation strengthens FLock’s influence in protocol design and connects our research with the global leaders shaping the architecture of Web3 and DeAI.
Community Engagement & Industry Events
- At the Web3 Consumer Meetup in Hong Kong, our CFO joined a high-impact panel to discuss how AI and crypto can move beyond hype. FLock’s position as the only team focused on decentralised federated learning demonstrated thought leadership in a space hungry for practical, user-first AI solutions. It also expanded our reach among developers, investors, and early adopters across Asia — a key region for Web3 growth.
- At the Hong Kong Web3 Festival, FLock’s CFO featured in a major panel on ‘The Convergence of AI Agents and Web3.’ Speaking alongside peers from GoKiteAI, GAIB AI, and Animoca Brands, we presented our approach to privacy-preserving autonomous ecosystems. These events helped position FLock as a foundational layer for the agent economy and significantly raised our profile among Web3 infrastructure builders and potential partners.
- Our presence at Token2049 Dubai allowed us to connect with some of the most influential funds, founders, and researchers in the DeAI and crypto ecosystem. Participating in such a high-profile global event helped cement FLock’s reputation as a serious infrastructure player — not just within niche AI circles but across the broader blockchain and investment communities.
- During ETH Seoul and Alibaba Cloud Day Korea, our team joined panels on scaling with AI alongside Univa and Upfall. These appearances highlighted how FLock’s infrastructure can power exponential growth in both decentralised and hybrid AI environments. Strengthening our relationship with Alibaba Cloud also supports future joint go-to-market strategies in Asia, where AI adoption is accelerating.
- Ahead of The Web Conference 2025, we hosted a fireside chat with over 50 leading AI and blockchain researchers. This session fostered valuable academic dialogue and seeded potential collaborations that will inform future iterations of FLock’s training infrastructure and governance. It also reinforced FLock’s commitment to building bridges between open-source communities and academic institutions.
- At Proof of Talk 2025 in France, our CEO shared the stage with top Web3 and AI minds to present FLock’s vision for privacy-preserving AI. By articulating a roadmap that includes tokenised incentives, federated training, and open governance, we made it clear that FLock is not just building infrastructure — we’re architecting a new economic model for autonomous AI. This raised our visibility among global decision-makers, founders, and capital allocators shaping the future of the AI economy.
- Our CEO met with U.S. lawmakers on Capitol Hill, helping to educate senators and representatives on the importance of decentralized, transparent AI infrastructure. These conversations are critical for shaping policy frameworks that support innovation while preserving digital rights — and they establish FLock as a credible voice in the regulatory future of AI.
- FLock was officially invited to the 100 Impact Leaders Dinner & Awards at the House of Lords in the UK Parliament. Attending this event alongside global leaders in blockchain and digital assets elevated FLock’s profile at the highest levels of government and industry. It signals growing recognition of our mission to democratise AI — not just within tech circles but in the halls of political power.
- Our VPs of Engineering and Growth attended a Base Ecosystem dinner in NYC hosted by Ribbit Capital and Base. This intimate, invite-only event gave us a chance to network with some of the top projects building on Base to discuss emerging trends, uncover collaboration opportunities, and understand how FLock’s privacy-preserving AI training platform can support the evolution of decentralised AI development on Base.
- At Consensus Toronto, we met with investors and existing partners as well as new potential partners. We had a number of meetings to further our efforts on a number of strategic fronts, especially around our Bittensor subnet (SN96), FLock OFF, and DEX listings.
