FLock.io and Qwen under Alibaba Cloud announce strategic partnership, Web3 AI needs to find complementary ecosystem position with Web2 AI
TechFlow Selected TechFlow Selected
FLock.io and Qwen under Alibaba Cloud announce strategic partnership, Web3 AI needs to find complementary ecosystem position with Web2 AI
Web3 AI urgently needs to find an ecological niche complementary to Web2 AI, addressing issues that centralized Web2 AI cannot solve—such as high computational costs, data privacy concerns, and fine-tuning models for vertical scenarios.
Navigating Web3 tides with focused insightsBy Haotian
Yesterday, @flock_io, a DeAI training platform in the Web3AI space, officially announced a collaboration with @Alibaba_Qwen, part of Alibaba Cloud. If I'm not mistaken, this marks the first time a Web2 AI entity has proactively initiated integration with Web3 AI. This partnership not only enables Flock to truly break out of its niche but also revitalizes the struggling Web3AI sector. Let me explain:
1) As I've mentioned in my pinned post, Web3 AI agents previously attempted to drive adoption through tokenomics and rapid deployment paradigms. Despite waves of FOMO-driven asset issuance, it became clear that when measured against Web2 AI on practicality and innovation, Web3 AI stood little chance.
Hence, breakthroughs in Web2 AI—such as Manus, MCP, and A2A—directly or indirectly burst the bubble around Web3 AI agents, leading to massive sell-offs in secondary markets.
2) So how do we break through? The path is actually quite clear: Web3 AI urgently needs to carve out a complementary niche alongside Web2 AI by addressing issues that centralized AI cannot solve—such as high computational costs, data privacy concerns, and fine-tuning for vertical-specific use cases.
The reason is simple: purely centralized AI models, after intense competition, will inevitably face bottlenecks in compute resource accessibility and cost, along with data privacy challenges. In contrast, Web3 AI’s distributed architecture can leverage idle computing power to reduce costs, employ privacy-preserving technologies like zero-knowledge proofs and TEEs, and promote model development and fine-tuning in vertical domains via data ownership and incentive mechanisms.
No matter the criticism, Web3 AI's decentralized architecture and flexible incentive models can deliver immediate solutions to certain persistent problems within Web2 AI.
3) Now, let’s discuss the collaboration between Flock and Qwen. Qwen is an open-source large language model developed by Alibaba Cloud. Thanks to its strong benchmark performance and the flexibility allowing developers to deploy and fine-tune locally, Qwen has become a popular choice among developers and research teams.
Flock, on the other hand, is a decentralized AI training platform integrating federated learning and distributed AI architecture. Its core feature is enabling distributed model training while keeping “data localized,” thereby protecting user privacy, ensuring transparent and traceable data contributions, and solving fine-tuning and application challenges in sensitive verticals like education and healthcare.
To be specific, Flock consists of three key components—let me briefly introduce them:
1. AI Arena: A competitive model training platform where users submit their models to compete with others on optimization outcomes and win rewards. Through gamified design, it incentivizes continuous local model refinement and helps identify superior base models;
2. FL Alliance (Federated Learning Alliance): Designed to address cross-organizational collaboration challenges in traditional sensitive sectors such as healthcare, education, and finance. The alliance enables multiple parties to collaboratively enhance model performance without sharing raw data, using localized training combined with a distributed coordination framework;
3. Moonbase: Serving as the nerve center of the Flock ecosystem, Moonbase functions as a decentralized model management and optimization platform. It offers various fine-tuning tools and computing power support (from providers of compute resources and data annotators), providing a distributed model repository while integrating fine-tuning tools, compute resources, and data annotation services to empower efficient local model optimization.
4) How should we interpret the collaboration between Qwen and Flock? In my view, the extended implications of this partnership far outweigh its immediate technical substance.
On one hand, amid the ongoing technological dominance of Web2 AI over Web3 AI, Qwen—representing tech giant Alibaba—already holds significant authority and influence in the AI community. Its decision to actively partner with a Web3 AI platform fully reflects recognition of Flock’s technical capabilities. Moreover, future joint research and development efforts between Flock and Qwen will deepen the interplay between Web3 AI and Web2 AI.
On the other hand, Web3 AI has long been criticized for having nothing more than a tokenomics shell, delivering poor real-world utility. Despite experimenting across numerous directions—including AI agents, platforms, and even frameworks—it has struggled to produce tangible solutions when applied to concrete areas like DeFai or Gamefai. This endorsement from a major Web2 tech player sets a precedent and provides directional clarity for the future of Web3 AI.
Most importantly, after a period dominated by FOMO-driven asset issuance, Web3 AI needs to regroup and refocus on delivering real results.
In truth, Web3 AI was never meant to be just an easier or faster way to deploy AI agents and issue tokens, nor a fundraising game. To thrive, it must seek cooperation with Web2 AI, complement each other’s ecological roles, and ultimately prove its indispensable value in this wave of AI advancement.
I’m genuinely encouraged to see more cross-boundary collaborations emerging between Web2 AI and Web3 AI.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
Haotian | CryptoInsight
