At the Web3 Summit in Berlin, SubQuery introduced their innovative decentralized AI inference hosting system.

Singapore, Singapore, August 21st, 2024, Chainwire

At the Web3 Summit in Berlin today, SubQuery In a significant reveal, SubQuery announced its cutting-edge decentralized AI inference hosting. During a live session, COO James Bayly highlighted the functionality of the latest LLama model running seamlessly across a decentralized network of Node Operators using SubQuery's internal testing environment.

SubQuery's mission is to enable developers to craft the future through decentralized technology. The organization is leading a movement towards creating the next generation of Web3 applications for a vast user base, prioritizing decentralization.

The SubQuery Network serves as the robust infrastructure that supports this initiative. Currently, it facilitates decentralized data indexers and RPCs, which are essential for developers working on decentralized applications (dApps). By establishing itself as a viable alternative to centralized platforms, SubQuery welcomes participation from anyone willing to be a node operator or delegator.

The impact of AI on various sectors, including Web3, is becoming more apparent. SubQuery has been keenly observing these advancements and has been quietly working to integrate AI functionalities into its decentralized framework. 'The Web3 Summit in Berlin is the ideal venue for us to launch this feature and showcase it live,' remarked James Bayly.

SubQuery concentrates on AI inference, which involves using pre-trained models to predict outcomes based on new information, rather than focusing on the training of those models. 'While there are commercial entities offering inference hosting for custom models, very few are present in the Web3 ecosystem,' James pointed out. 'Our decentralized network is perfectly positioned for reliable and sustainable AI model hosting.'

Presently, the AI inference landscape is largely controlled by major centralized cloud providers that impose hefty charges and frequently utilize clients' data to refine their proprietary models. 'Companies like OpenAI and Google Cloud AI not only charge premium prices but also exploit your data to enhance their closed-source models,' James explained. SubQuery aims to present a cost-effective, open-source alternative for hosting AI models in production. 'We aspire to enable users to launch a production-ready LLM model via our network in just ten minutes,' he further added.

'Depending on closed-source AI solutions risks consolidating power among a few dominant firms, perpetuating their influence,' James cautioned. 'By utilizing a decentralized network for AI inference, we prevent any single entity from controlling or misusing user data. Prompts are shared across hundreds of node operators, ensuring privacy and fostering an open-source ecosystem.'

The SubQuery Network is set to deliver cutting-edge hosting for the latest open-source AI models, paving the way for scalable and user-friendly AI services tailored for Web3. By adopting a community-based approach, SubQuery promotes extensive decentralized AI inference, thus empowering a varied network of independent Node Operators.

About SubQuery

SubQuery Our network is redefining web3 infrastructure with tools that equip builders to decentralize the future without compromise. Our adaptable DePIN infrastructure network hosts the quickest data indexers, the most scalable RPCs, innovative Data Nodes, and top-tier open-source AI models. We are the foundation of the web3 ecosystem, enabling blockchain developers and their state-of-the-art applications to thrive. We don’t just see ourselves as a business; we are part of a movement towards an inclusive and decentralized web3. Let’s collaborate to mold the future of web3.

​​​​ Linktree Website Discord Telegram Twitter Blog Medium LinkedIn YouTube

Contact

Head of Marketing
Brittany Seales
SubQuery PTE LTD