TLDR:
- The convergence of AI and Crypto, known as AI x Crypto, combines artificial intelligence with blockchain technology. This integration aims to enhance transparency, security, and user control in AI applications through decentralization, emerging as one of the most compelling innovations in the current tech landscape.
- In the new asset-based landscape of AI x Crypto, the evolution of computing power, models, and data takes center stage. Computing power assetization focuses on decentralized computing and AI inference, while model/Agent assetization drives the NFTization of virtual characters. Data assetization emphasizes the liberation and utilization of private data resources. However, this field still faces challenges such as immature business models and contradictions between interdisciplinary expertise and practitioner preferences.
- Despite the pain points in the AI x Crypto space, its importance cannot be overstated. AI is seen as a key force driving the next technological revolution, and decentralized computing power is expected to further unlock AI’s potential. Computing power plays a crucial role in projects, serving as a significant metric for technical strength and market consensus, and is expected to facilitate innovative resource allocation and value distribution.
The highly anticipated NVIDIA GPU Technology Conference 2024 (NVIDIA GTC) will take place in San Jose, California, from March 17–21. Jensen Huang, the “godfather of AI” and CEO of NVIDIA, will deliver a keynote speech titled “Don’t Miss the AI Revolution.” He will share the latest achievements and practices of NVIDIA in hardware and software co-optimization for Large Language Models (LLM), which will profoundly impact and advance the integration of AI and blockchain technology. NVIDIA will showcase its full-stack LLM solution, covering architecture design, FP8 training, multi-precision training, and efficient inference deployment. Through core technologies such as Megatron-Core, TensorRT-LLM, and Triton Inference Server, NVIDIA empowers developers to build high-performance, stable, and reliable large model application environments in data center architectures.
In the field of Crypto AI, NVIDIA’s technological advancements have presented new growth opportunities for projects across the application, middleware, and infrastructure layers. Application-layer projects can leverage optimized blockchain performance and the flexibility of smart contracts to create richer and more efficient Crypto AI applications. Middleware-layer projects can innovate with the support of NVIDIA’s technology, developing permissionless and autonomous Crypto AI Agents, as well as on-chain machine learning and model hosting projects that ensure data security and transparency in the model training process. Infrastructure-layer projects can draw inspiration from NVIDIA’s breakthroughs to build safer and more efficient AI public chains and decentralized computing networks.
At the blockchain application level, AI technology will deepen the flexibility and intelligence of smart contracts, enhancing productivity scenarios and user experiences in entertainment settings. This conference will once again bring global attention to the latest developments in AI, with various related AI crypto assets experiencing price increases in advance. Notable projects such as Worldcoin, Bittensor, and Render have reached new highs in their respective token prices. Undoubtedly, AI-related assets will be a significant focus of speculation and investment throughout 2024. This article will provide an overview of notable AI crypto projects to watch in the first quarter of 2024.
In terms of generative AI trends, large models, as the core of the AI industry chain, are evolving towards multi-modality and AI Agents to better adapt to productivity and entertainment applications. In productivity scenarios, AI demonstrates its business value by achieving cost reduction and efficiency enhancement. In entertainment settings, AI enhances user satisfaction by improving interactive experiences. Similar innovations are emerging in different scenarios and at different levels within the Crypto AI space. This article categorizes crypto projects into three main layers: application, middleware (including Agent, model hosting, data, and privacy authentication subcategories), and infrastructure (including public chains and decentralized computing networks).
Crypto AI application-layer projects are characterized by their use of AI algorithms to enhance the intelligence and autonomy of blockchains. By introducing machine learning, deep learning, and other AI techniques, these projects not only optimize blockchain performance but also make smart contracts more flexible and intelligent.
Middleware-layer projects, serving as a bridge between the application and infrastructure layers, are beginning to exhibit diverse innovations. Some focus on permissionless and autonomous Crypto AI Agents, while others involve on-chain machine learning and model hosting. By storing model training processes and data on the blockchain, these projects ensure data security, immutability, and transparency in the model training process. Data-focused projects aim to address issues related to data storage, processing, and analysis on the blockchain, leveraging distributed storage and encryption techniques to ensure data security, privacy, and efficient processing.
Infrastructure-layer projects include innovative AI public chains like Bittensor, which store model training processes and data on the blockchain, ensuring data security, immutability, and transparency. Additionally, there is a growing number of decentralized computing network projects that combine AI with popular narratives in the Depin space. Coupled with the rising stock prices of AI computing power companies like NVIDIA, these projects have garnered significant market attention.
The cutting-edge AI technologies and solutions presented at the 2024 NVIDIA GTC conference are expected to ignite a new wave of excitement around AI crypto projects in the Web3 world. From generative AI applications to on-chain machine learning, middleware and Agent innovations, and the emergence of public chains and decentralized computing networks, Crypto AI is poised to be a hot crypto sector in 2024 and beyond.
- NFPrompt: Co-created by seasoned Web3 builders and accomplished contemporary artists, NFPrompt is the first AI-driven user-generated content (UGC) platform in the Web3 space, dedicated to enabling users to easily create AI artworks and mint them into NFTs. NFPrompt employs a dual-token model consisting of NFP and cNFP native tokens. NFP tokens are utilized for community governance, staking rewards, and royalty sharing, while cNFP serves as a reward system where users can earn cNFP through daily activities and community participation, potentially receiving NFP token airdrops and other incentives.
- QnA3: The core strength of QnA3 lies in its unique “AI+Research” and “AI+Trading” dual-wheel drive model. By leveraging advanced RAG (Retrieval Augmented Generation) technology, QnA3 swiftly and accurately acquires professional knowledge in the Web3 domain. Combined with the powerful reasoning capabilities of LLMs (Large Language Models), it provides users with efficient and precise information management and trading assistance. Additionally, QnA3 explores the new paradigm of “AI+DePIN,” utilizing decentralized physical device computing power for AI model training. This offers users passive income opportunities while facilitating the deep integration of Web3 with the real economy. QnA3’s development path begins with the launch of a question-and-answer function to attract initial users, followed by the introduction of Telegram bots and data mining features to achieve rapid user growth. Future plans include the rollout of innovative products like hardware wallets and desktop robots to cater to users’ full-lifecycle needs across various scenarios.
- ChainGPT: ChainGPT’s technological highlight is its transformer-based language model, which combines deep learning with extensive data training to ensure accurate and rapid predictions. Its highly scalable AI model can process vast amounts of data and quickly provide solutions to various cryptographic and blockchain issues. Moreover, ChainGPT is a comprehensive ecosystem that integrates blockchain technology, such as using $CGPT tokens for model access, staking, and mining mechanisms to incentivize user contributions to the network. The leadership team comprises Gintare Kairyte, a professional with seven years of experience in Web3 and blockchain; Amid Yazdi, a senior civil engineer and business mentor with years of experience in Web3 projects; and Marc Constanti, a technical expert from Barcelona who has held leadership positions at renowned companies like Sony Music. Together, they form a powerful team driving ChainGPT’s success in the blockchain domain.
- Adot: Adot is a Web3 decentralized intelligent search engine dedicated to providing users and developers with comprehensive coverage, acquisition, aggregation, and indexing services for on-chain data. Its functionality encompasses various search methods like tag search, full-text search, and structured parsing, while supporting customized conditional searches and API data calling interfaces. Additionally, Adot offers AI search, content AI summarization, and sharing features, enhancing the user experience. In terms of architecture, Adot integrates artificial intelligence technology for modeling, captures data sources from various platforms through a hybrid data approach, undergoes structural parsing, AI modeling, and structured indexing for data cleaning and processing, and finally presents the end results through search algorithms in the application layer. This design allows Adot to achieve on-chain data search, AI summarization, AI sharing, and other high-quality content output. The key highlight of Adot is its integration of AI technology to facilitate on-chain massive data retrieval and data combing, along with the provision of rich AI application components.
- Sleepless AI: Sleepless AI’s HIM is a Web3 female-oriented anime-style virtual boyfriend game that offers players a unique love experience. In the game, virtual boyfriends are stored on the blockchain in the form of SBTs, which include NFTs for clothing, accessories, rooms, and more. These SBTs constitute the non-fungible game characters that players cultivate, storing basic information, nine-dimensional personality traits, and nine-dimensional relationship attributes. They also store important memories, conversations, emotional inscriptions, and other content related to the virtual boyfriend. Additionally, players can regularly receive treasure chests that unlock limited-edition NFT clothing or room items through airdrops. The game features various gameplay options such as card drawing, dressing up, conversation, collection, adventure battles, cooking, traveling, and NFT trading. The core of the HIM project is the integration of AI technology with cryptocurrency, creating a Web3 game with rich interactive experiences, high-quality gameplay, and artistic content.
- KaitoAI: Kaito is a Web3 search engine company based on AI large language models, dedicated to solving the problem of information retrieval in Web3 that traditional search engines cannot address. Founded in March 2022, Kaito’s CEO and founder, Yu, focuses on leading the team to make actionable insights in blockchain networks more accessible through AI technology. Yunzhong, Kaito’s co-founder, previously led an AI team responsible for building Meta’s search and recommendation engine and now plays a crucial role in Kaito’s AI research platform, which provides searchable insights for digital assets. The company uses Auto-GPT technology and powerful GPT4 large language models, with plans to introduce state-of-the-art AI technology into a broader range of Web3 infrastructure and applications, including data labeling based on large language models, privacy-preserving machine learning training, content distribution, fact-checking, and transaction optimization.
- AwesomeQA: AwesomeQA is an AI-powered customer support solution specifically designed for Web3. The company has provided customer support to over 60 Web3 companies and projects. AwesomeQA’s CEO, Alexander, believes in the power of decentralized communities to drive societal technological progress. CTO Korbinian views strong product communities as the foundation of trust and sees AI as a catalyst for human-centered technological development. Sales and Marketing Manager Syed emphasizes the importance of brand community building, while Product Designer Miguel prioritizes a user-centered design approach. Software Engineer Aykut utilizes AI to optimize community management and enhance the user experience.
- X23: X23.ai is a service platform focused on simplifying Web3 data. It utilizes advanced AI technology to transform diverse and complex Web3 information, such as GitHub pull requests and governance discussions, into clear and concise summaries. X23.ai also offers real-time alerts, keeping users informed of the latest Web3 developments through Telegram, Slack, email, and other means.
- Scopechat: Scopechat is an easy-to-use and intuitive AI assistant designed for cryptocurrency traders and users of all levels. Leveraging 0xScope’s extensive coverage, high-quality data warehouse, AI-driven algorithms, and large language models (LLMs), Scopechat provides accurate and comprehensive answers to Web3-related questions using both public and proprietary datasets. It allows users to obtain real-time Web3 data and insights from a single AI assistant, eliminating the need to switch between multiple applications to check relevant information. The core team behind 0xScope includes co-founder Pedro Torres, co-founder and CTO Colin Yu, and CMO RAGINI Raffaele.
- Numerai: Numerai envisions revolutionizing investment and fund management on Wall Street through artificial intelligence and collaborative resources. The company utilizes a token called NMR to incentivize data scientists to predict market trends using machine learning algorithms, aiming to develop highly accurate market strategies. Data scientists can participate in weekly machine learning competitions hosted by Numerai to earn NMR tokens by submitting prediction strategies. These competitions use historical market data, and participants are rewarded with NMR if their predictions are accurate. Additionally, the NMR token has a staking mechanism that allows strategy providers to stake their NMR tokens to have their strategies applied in real-market funds, earning profits or incurring losses based on actual performance.
- Noya.ai: Noya.ai is a DeFi application that enables AI models to exist entirely on-chain, with ZKML technology used to verify model outputs in zksnark circuits. It employs Halo2 and ezkl libraries to achieve efficient operations and secure proofs. The core component of Noya.ai is Omnivaults, an automated yield generation strategy tool that utilizes AI models to manage and optimize assets from different chains.
- RoboNet: RoboNet is an open database aimed at sharing robot experiences and providing initial video frame pools for learning generalizable models in vision-based robotic manipulation. It comprises over 15 million video frames covering seven different robot platforms. RoboNet’s research focuses on exploring how this dataset can be used to learn models that can effectively work with new objects, tasks, scenes, camera viewpoints, grippers, and even entirely new robots. The database combines two different learning algorithms: visual foresight and supervised inverse models.
- Hera: Hera is an AI-driven, multi-chain decentralized exchange (DEX) aggregator designed to provide users with the best trading experience. It analyzes trading volume, prices, and liquidity to identify the optimal and most profitable trading paths on decentralized exchanges. This means that regardless of the token pair a user wants to trade, Hera can help them allocate their trades to multiple liquidity sources in the most optimized way, maximizing trading efficiency and minimizing costs.
The team behind Hera includes Yasin as the co-founder responsible for technical direction and innovation, Vedat focusing on product technology research and development, Christian as the head of business development driving the company’s growth and market expansion. As the product owner, Sedat leads the design and optimization of the product. Madeline is the marketing assistant, providing strong support for brand communication. Richard is the community leader, while Victory, as the community manager, is responsible for the daily operation and management of the community. Kathrin plays a dual role as a content creator and researcher, providing the team with rich content and in-depth market insights.
- MyShell: MyShell is a no-code robot creation platform aimed at lowering the programming barrier, enabling users without a programming background to easily create personalized robots. MyShell offers a wide range of robots, including language learning, education, and utility types, catering to different user needs. MyShell has introduced the Robot Workshop feature, with nearly 60 user-created robots currently available and over 100 private robots. Users can combine robots based on their interests, while robot authors can select high-quality models, and model authors can gain access to application scenarios and high-quality data on the platform. MyShell believes that large language models should serve as the super glue connecting other modalities and services, playing a coordinating role and enabling different small models to complete complex tasks together.
The founders of MyShell, Rick and Ethan, have extensive experience in AI entrepreneurship. Rick has founded companies in the fields of graphics and AR since 2013, and Ethan, a former VR startup founder, later joined an AI unicorn company to lead the robotics department. Leveraging the power of large models like GPT-4 and a multimodal approach, they have chosen to create innovative software robot products.
2. PAAL AI: PAAL provides users with a customized and scalable artificial intelligence experience. The platform is built on AI technology, integrating natural language processing (NLP), machine learning (ML), and automation features, and is fused with blockchain technology.
3. Autonolas: Autonolas is an innovative project that combines artificial intelligence with blockchain technology. It has constructed a set of smart contract protocols, resembling an app store, allowing developers to register and monetize their AI-driven services. These services can economically interact and have their rules collectively decided by DAO members. Through Autonolas, AI Agents can significantly improve the performance of various services such as managing liquidity pools, predicting markets, asset acquisition, oracles, and cross-chain bridges, while achieving collaboration and standardization among services through Ensemble AI and networked AI.
Developers can register their code as NFTs and receive corresponding rewards through the use of autonomous services if adopted by the DAO. Additionally, Autonolas utilizes a liquidity growth mechanism owned by the protocol, which pairs code with capital using a bonding mechanism, incentivizing OLAS token holders to provide liquidity and earn staking rewards. The OLAS token not only facilitates DAO governance but also offers staking privileges, participation in liquidity mining, and other functions, supporting the project’s long-term growth.
4. MorpheusAI: To enable more users to access and utilize intelligent Agents and enhance the decentralization of their infrastructure, MorpheusAI has proposed the development of the Morpheus Network. This network includes a fairly launched token, MOR, which incentivizes four key contributors within the community: community members building interfaces, programmers contributing to Morpheus software/Agents, capital providers offering liquidity, and participants providing computing, storage, and bandwidth resources. MorpheusAI aims to provide these API interfaces and decentralized cloud capabilities by launching the network and issuing tokens, rewarding those who contribute to the public blockchain infrastructure for the intelligent Agent community.
5. Arbius: Arbius is a blockchain-based decentralized machine learning platform. It allows users to interact using its native token, AIUS, for various activities such as generating artworks, participating in liquidity provision to earn staking rewards, and mining. Users can interact on the Arbitrum Nova chain through Ethereum-compatible wallets like Metamask, experiencing image generation by selecting a model and entering creative descriptions in the prompt box. The generated results are recorded on-chain for everyone to view. Users seeking passive income can purchase AIUS and combine it with other assets to form liquidity pools, which can then be staked on the Arbius platform to earn newly allocated AIUS rewards.
The Arbius.ai team comprises experts from various fields: Charlie handles social media and has extensive experience collaborating with crypto companies; J Master Pig is a product leader and former TV producer who has worked for multiple international corporations; Kasumi is the founder of MistCoin and has profound insights into economics and future currencies; Mar is a graphic designer familiar with crypto art; Mary Doble specializes in full-stack digital marketing and has worked for large international enterprises; Oscar Salas is an expert in marketing and video animation who has produced content for several well-known crypto companies.
6. CharacterX: CharacterX is an entertainment social network that combines decentralized AI technology with the advantages of cryptocurrencies. It focuses on building an immersive AI experience environment where users can make friends, entertain themselves, learn, and even earn rewards through interaction. Unlike general-purpose AI conversation tools, CharacterX emphasizes emotional companionship and entertainment, allowing creators to create unique AI avatars based on real people or literary characters and enhance user experience using multisensory technologies such as images, voice, 3D, and AR.
The founding members of CharacterX include Rene and Jeremy, both alumni of Stanford University. Most of the other team members have graduated from prestigious universities like the University of Chicago and Peking University or have worked at leading tech companies like TikTok, Tencent, and Facebook, possessing technical expertise and resources in areas such as crypto development, AI applications, and 3D modeling.
7. Fetch.ai: Fetch.AI is a company that deeply integrates blockchain and artificial intelligence technology. Founded by Humayun Sheikh, Thomas Hain, and Toby Simpson, its core technologies include the Fetch.AI mainchain and the Autonomous Economic Agent (AEA) architecture. As intelligent Agents, AEAs enable efficient, secure, and intelligent transactions in various fields like logistics and finance through autonomous decision-making, collaboration, and learning. The AEA architecture comprises four key modules: AEA Agent, AEA communication, AEA skills, and AEA protocols, which work together to enable Agents to interact and optimize strategies automatically in a decentralized environment.
8. Libertai: The Libertai project is a decentralized artificial intelligence assistant that utilizes advanced machine learning algorithms, specializing in enhancing language processing capabilities. Its large-scale language model leverages the advantages of decentralized technology, combining IPFS and aleph.im to construct a computing network.
9. ChainML: ChainML is a decentralized computing platform that has released an open-source AI Agent platform called Council as its core highlight. Council integrates with multiple LLMs, including GPT-4, enabling enterprises to achieve innovative generative AI applications. In the short term, ChainML drives demand for AI on the decentralized computing network through Council. In the long term, it aims to build a comprehensive AI+web3 platform, including model and computing marketplaces. Its tool library is compatible with various complex data types, providing features like IPFS data access, ML service monitoring, and troubleshooting, while integrating with developer tools like Alchemy and Truffle. ChainML ensures the validity of model training and inference through innovative use of Proof of Learning, Proof of Inference, and Proof of Quality.
10. Shinkai: The Shinkai project aims to break the limitations of existing vector databases by developing advanced technologies like Vector Resources and VectorFS, providing AI Agents with a more flexible, versatile, and easily interactable data format. Through its proprietary Shinkai Node, AI-specific file system VectorFS, and decentralized P2P network, Shinkai significantly enhances AI’s ability to handle daily tasks and real-time access to the latest global information while ensuring user data privacy. Users can seamlessly access Shinkai through Chrome, web browsers, or mobile devices, integrating AI with common tools like email, Google Docs, Slack, etc., intelligently managing various documents and notes and keeping information synchronized and up-to-date.
- Giza: Giza is an AI-based encryption project that has completed a $3 million Pre-Seed funding round led by CoinFund, with participation from StarkWare, TA Ventures, and Arrington Capital. The project aims to introduce artificial intelligence into the blockchain space and achieve blockchain transaction verification with less computing power by integrating zero-knowledge proof technology with AI.
- Inference Labs: Inference Labs is building advanced AI infrastructure and products to enable decentralized, transparent, and secure AI applications. By constructing AI inference interoperability on the blockchain, the lab eliminates the need for trust in third parties, allowing global users to access AI without risk.
- Axiom: Axiom.ai helps users create automated robots quickly and easily through a browser plugin, which can perform tasks such as data scraping, data entry, and other interface interactions (e.g., clicking, scrolling, entering text) on any website. It offers easy installation, supports custom configuration and template editing to meet individual needs. The tool can be run manually or set up as scheduled tasks or triggered by external events through Zapier integration.
- EZKL: EZKL is an open-source tool dedicated to simplifying the construction and application of zero-knowledge proofs, particularly for AI/ML models and computational graph programs built on frameworks like PyTorch. With EZKL, developers can describe their programs from a high-level perspective and generate corresponding zero-knowledge provers and verifiers, enabling efficient and secure proofs about the execution results of neural networks (whether public or private models) on different data (public or private). EZKL supports Python API calls, provides command-line tools and WebAssembly support, and can be flexibly applied to various complex off-chain computation scenarios without limiting input size or relying on centralized coordinators. Its lifecycle includes three stages: setup, proving, and verification, corresponding to defining the proof structure, generating the proof, and verifying the proof, respectively. Developers can easily integrate EZKL into their projects through pip installation, leveraging zero-knowledge proof technology to protect privacy while ensuring the validity of computation results.
- Flock: Flock enables developers to collaboratively build AI models on a full stack and provides fair rewards to users who contribute data. It allows users to optimize large language models (such as LLMs, Stable Diffusion, etc.) while maintaining data locality and privacy and achieves fast and energy-efficient fine-tuning through LoRA technology. The entire process is governed on-chain, and computation, training, and model hosting are decentralized, ensuring that all participants (including data, feedback, and computing power contributors) share the benefits. Flock has a wide range of application scenarios, including on-chain GPT models, real-time Web3 insights, programming assistance, gaming AI companions, private health monitoring, anonymous and reliable DeFi credit scoring systems, and user privacy-focused intelligent auto-completion tools.
- Spectral: Spectral AI is a Dallas-based medical diagnostic predictive AI company specializing in revolutionizing wound care management through its DeepView® system, particularly for burn and diabetic foot ulcer patients. DeepView® is an algorithm-based predictive diagnostic device that can objectively assess wound healing potential immediately before intervention, aiming to exceed current standards of care and improve patient outcomes with faster and more accurate treatment insights while reducing healthcare costs.
- PublicAI: PublicAI is an innovative platform that combines Web3 technology with AI data annotation. By constructing a decentralized global collaboration network, it integrates data processing with distributed economic incentives, addressing the demand for high-quality vertical data in AI model training. The platform leverages blockchain technology to ensure the credibility, efficiency, and fairness of data processing, allowing professionals worldwide to directly participate in data annotation tasks and earn token rewards. PublicAI reduces intermediaries, achieving fair revenue distribution between annotators and AI research institutions while lowering costs and improving data quality and annotation efficiency.
- Rainfall: Rainfall is an innovative project that combines cutting-edge edge AI technology with blockchain to create a decentralized, privacy-focused personal intelligence platform. The platform focuses on enabling users to safely control their data and derive fair economic value from it. By analyzing and contextualizing user-authorized data, it generates high-value anonymous social intelligence insights in real-time. These insights are sold anonymously to businesses to improve customer service and attract new customers, while an integrated fair value pricing engine automates the transaction process, ensuring users receive their due rewards from personal data without sacrificing privacy.
- Origin Trail: OriginTrail is an ecosystem focused on trusted data sharing, leveraging blockchain technology to create a decentralized knowledge graph infrastructure. The project’s core objective is to address issues of trustworthiness and accuracy in AI by incentivizing community members to contribute and validate high-quality knowledge assets/IP through its native token, TRAC, while protecting data owners’ rights. The economic benefits of TRAC flow back to the ecosystem’s node participants, resulting in significant growth in both knowledge assets and node revenue as the project develops.
- Grass: Grass is a decentralized web scraping platform that utilizes idle home network bandwidth from global users to scrape and validate public web data for AI projects, with a total funding of $4.5 million. Grass plans to achieve 24/7 operation through an upcoming Android mobile application, significantly increasing the number of active nodes in the network and accelerating the project’s progress towards reaching the IP address threshold required to trigger compensation mechanisms. Additionally, the application will enhance the referral mechanism, allowing users to earn continuous commission rewards by referring others, further driving network growth and expansion.
- Bagel Network: Bagel Network is an innovative AI data marketplace designed to facilitate collaboration between machine learning engineers, researchers, and AI agents. The platform supports the secure building, trading, and licensing of datasets while prioritizing privacy and promoting responsible data evolution through verifiable data integrity. It fosters the integration of humans and AI, emphasizing a symbiotic relationship between the two.
- Sapien: Sapien emerged from an engineering and innovation team in San Francisco with a track record of developing chatbots for in-flight use by major airlines and creating iOS/Android voice AI SDKs for the NFL and NBA. In the summer of 2021, Sapien was invited to participate in OpenAI’s developer beta program, where they deeply explored cutting-edge tools such as ChatGTP, Whisper, and Dalle.
- DataOS: DataOS is built upon an open data lake on the blockchain and operates within a decentralized incentive framework through an autonomous agent network. It transforms applications into adaptable and continuously evolving entities that meet users’ changing needs. Additionally, DataOS breaks down content format limitations through technologies like language models, enabling the universal composition of digital content.
- Fraction AI: Fraction AI is designed specifically for real-world AI demands. It powers global data scientists, machine learning engineers, researchers, and AI innovators by improving data pipelines and accelerating machine learning speeds. Fraction AI leverages real user experiences to develop high-quality machine learning training data, achieving supercharged enhancements in artificial intelligence. It serves as a powerful ally in the AI field, providing solid support for AI innovation and application.
- Worldcoin: Worldcoin is a project dedicated to building a globally inclusive identity authentication and financial services network, aiming to issue a digital currency called WLD based on Proof of Personhood. The core components of the Worldcoin project, World ID and WLD, complement each other and utilize customized biometric hardware to verify the authenticity and uniqueness of individuals globally, addressing the increasing challenge of distinguishing between human and AI online identities. The project’s vision is to promote global democratic participation, economic growth, and potentially AI-funded universal basic income while ensuring user privacy. Worldcoin addresses the missing digital element of personal identification and collaborates with Ethereum to ensure network security.
- Nevermined: Nevermined is a platform centered around digital assets, enabling the creation and operation of a digital ecosystem for interactions between various entities and digital assets. In Nevermined, assets consist of three components: unique on-chain asset registration information (including the asset’s decentralized identifier DID and a reference to the asset metadata DDO), metadata describing the asset’s details, and all files associated with the asset. It’s worth noting that Nevermined does not store the asset-related files but retains the asset’s DID and metadata to facilitate asset discovery.
- Numbers: Numbers Station is a conversational AI-driven enterprise-level data analytics platform developed by a Stanford research team. It enables data analysis, text matching, and API access through conversation, eliminating the need for specialized technical knowledge or external systems. It removes the barriers to understanding databases, schemas, and models, leveraging decades of AI research to solve real-world business problems and enhance data accessibility. Numbers Station allows any user to safely and economically extract insights from existing data using AI.
- ZAMA: Zama is an open-source cryptography company specializing in building state-of-the-art fully homomorphic encryption (FHE) solutions for blockchain and artificial intelligence. Its products include TFHE-rs (a pure Rust implementation of the TFHE library for boolean and small integer operations on encrypted data), Concrete (a TFHE compiler that converts Python programs into FHE equivalents), Concrete ML (a privacy-preserving machine learning framework based on Concrete that can be integrated with traditional ML frameworks), and fhEVM (an implementation of private smart contracts on the EVM using homomorphic encryption).
- PRIVASEA: Privasea is an innovative AI network that utilizes fully homomorphic encryption technology and blockchain incentives to address data privacy issues and meet the demands of collaborative AI. The network manages Gas fees and miner proof-of-stake through smart contracts and token incentives, enabling secure and efficient processing. Privasea’s core value lies in providing confidential machine learning inference and data value sharing, promoting knowledge exchange and collaboration while reducing privacy risks.
- zkHoldem: A fair and fully on-chain Texas Hold’em platform powered by zkSNARK technology, ensuring absolute game integrity.
- Modulus Labs: Modulus Labs has chosen zero-knowledge proofs (ZKP), a hot topic in the cryptography field in recent years, and combined it with machine learning to create zkML (Zero-Knowledge Machine Learning). This fusion technology allows external observers to verify the correct execution of AI algorithms without revealing their internal complexities, enhancing public trust in AI decisions.
- Bittensor: Bittensor is a decentralized machine learning protocol that combines blockchain technology and incentive mechanisms to create a peer-to-peer marketplace for trading machine intelligence. It enables global machine learning nodes to collaboratively train and learn specific problems, thereby enhancing overall intelligence. Its native token, TAO, serves as the network’s reward and access token, mined through a Proof-of-Work (POW) mechanism, and incentivizes participants to contribute computing resources, expertise, and innovation through a token economic model. The Bittensor network comprises nodes (miners and validators), where miners provide locally hosted machine learning services, and validators ensure data quality and model integrity. The network employs a Shapley value-based scoring method to distribute rewards based on each model’s marginal contribution to the network’s predictive performance. With a total supply of 21 million TAO tokens, it follows a halving mechanism similar to Bitcoin.
Bittensor is a blockchain project created by AI researchers aimed at incentivizing global machine learning nodes through cryptocurrency, promoting the decentralization of AI development. It introduces innovative mechanisms such as distributed expert models and proof-of-intelligence to reward useful machine learning models and outcomes, fostering the growth of a decentralized AI ecosystem. Bittensor’s architectural design reflects its pursuit of building a robust AI ecosystem, including a layered structure of miners, validators, enterprises, and consumers, supporting AI innovation in all aspects.
Its key participants, including miners and validators, are encouraged to compete and improve model performance through incentive mechanisms. Bittensor seeks to achieve composite and decentralized intelligence through a subnetwork model and employs a token economic model influenced by Bitcoin, aiming to promote the democratization and decentralized iterative learning of AI technology. Compared to traditional centralized AI models, Bittensor facilitates the openness and sharing of AI technology, potentially accelerating technological progress, reducing application costs, and enabling more individuals and small businesses to participate in AI innovation.
2. AgentLayer: AgentLayer is a blockchain-based L2 public chain specifically designed for AI Agents. Its objective is to establish a trusted, secure, and efficient decentralized infrastructure for AI Agent collaboration, encompassing direct cooperation among multiple Agents, as well as coordination between Agents and upper-layer applications, hosted models, and the underlying decentralized computing networks. AgentLayer provides the necessary infrastructure and tools for the development, deployment, trading, and collaboration of AI Agents. Additionally, it offers support for the financialization of AI Agents through its AgentFi module. The ultimate goal of AgentLayer is to construct an open and interoperable public chain centered around decentralized AI Agents and serve as a platform for the issuance of AI capital.
3. Hyperspace: Hyperspace is a platform dedicated to calculating and displaying the rarity of NFTs. It draws inspiration and builds upon the pioneering work of projects like rarity.tools while incorporating its own unique algorithms. The platform computes the rarity of each individual characteristic for every NFT in a collection, aggregates the rarity scores for all characteristics of an NFT, and uses these scores to rank the entire collection. The NFT with the highest rarity score occupies the top position. In cases where multiple NFTs have the same rarity score, they are assigned the same rank, and subsequent ranks are adjusted accordingly. It’s worth noting that Hyperspace excludes numerical range attributes from its calculations and currently does not dynamically update rarity as NFTs are destroyed or added.
4. Heurist: Heuristica leverages generative AI technology and mind mapping visualization to offer highly personalized and self-directed learning experiences. It caters not only to individual users expanding their knowledge but also serves as an educational tool for teachers to facilitate personalized instruction, project-based learning, differentiated teaching, and student assessment. Heuristica’s ability to learn and optimize its services based on user interaction makes it akin to an intelligent library that evolves with users’ interests, significantly advancing the development of personalized autonomous learning.
5. Gensyn: Gensyn utilizes the Substripe protocol to build a first-layer Proof-of-Stake blockchain. Through smart contract mechanisms, it effectively distributes and rewards machine learning tasks, significantly reducing the cost of deep learning training. Its verification system incorporates innovative designs such as probabilistic proof-of-learning, graph-based localization protocols, and Truebit-style incentive games to ensure the correct execution of computational tasks and fair transactions through token incentives. The system comprises four types of roles: submitters, solvers, verifiers, and whistleblowers, who collaborate to complete large-scale neural network training verification in a trustless environment, with verification costs maintaining a linear relationship with model size.
6. FedML: FedML is an open-source library and MLOps platform committed to providing simple yet versatile APIs for running machine learning anywhere and at any scale, with a particular focus on supporting federated learning and distributed training. The FedML MLOps platform further simplifies the practical application process of federated learning, enabling users to perform distributed machine learning securely with zero code and cross-platform capabilities, ensuring data remains local and maximizing privacy and efficiency. Through FedML, users can easily deploy and execute federated learning projects across data silos, smart devices, and organizations, achieving a seamless transition from theoretical research to practical applications.
7. Ritual: Ritual is a decentralized AI network project aimed at supporting the Ritual Chain and Infernet by building the first AI-native AVS (Artificial Intelligence Verification System) to enhance security and decentralization. The project integrates AI coprocessors to embed AI-native computing capabilities into the blockchain, addressing issues such as computational integrity, privacy protection, and censorship resistance, and promoting the decentralized development of AI technology.
Ritual has announced a partnership with the restaking protocol EigenLayer, allowing EigenLayer operators to access and utilize novel revenue streams. Additionally, Ritual successfully raised $25 million in funding led by Archetype in November 2023, with participation from Accomplice and Robot Ventures. Ritual has engaged BitMEX co-founder Arthur Hayes as a project advisor to assist in advancing the financialization of AI models, GPUs, and data.
8. Vanna: Vanna.ai is an AI assistant powered by natural language processing technology. It enables users to interact with SQL databases through everyday conversational formats, eliminating the need to write SQL code to retrieve information. Vanna.ai employs a Retrieval-Augmented Generation (RAG) model that, once trained, accurately converts user queries in natural language into SQL queries and executes them in a local environment.
The tool supports various database types, allows users to customize training on their own data to improve query accuracy, and provides flexible interface options, including Jupyter Notebook, web applications, and Slack bots. Vanna.ai emphasizes data security and privacy, ensuring that all query operations are performed locally, and it has self-learning and optimization capabilities that enhance query precision over time.
- Akash: Akash is a decentralized cloud computing marketplace that utilizes open networks to build a secure and efficient computing resource trading platform. Users can deploy complex applications through the SDL-driven orchestration layer on this platform, leverage Kubernetes-based architecture to ensure the safe and stable operation of applications, and provide features such as unlimited storage and dedicated IP leasing. Akash boasts characteristics of permissionlessness, peer-to-peer communication, and privacy, ensuring data security and resistance to censorship. It provides a decentralized, reliable, and democratized application deployment environment for all innovators. Users can easily deploy applications with the help of Cloudmos tools or monetize their cloud resources by joining the Akash marketplace. Whether you’re a developer deploying applications or a cloud resource owner becoming a provider, you can quickly get started and start using the platform by referring to Akash documentation.The Akash Network leverages globally underutilized GPU resources to create an efficient market for resource owners and demanders. Analogous to Airbnb’s business model, Akash enables GPU owners to rent out their computing power, lowering the entry barrier for AI and machine learning fields. Its native token, AKT, plays a crucial role in the network, used for paying resource fees, participating in network governance, and incentivizing user participation. Through incentive mechanisms such as token rewards and transaction fees, Akash encourages more users to provide resources and ensures the healthy development of the network.
- Clore: Clore.ai is a distributed supercomputer service platform that aggregates GPU computing resources from global community members, enabling users to rent these high-performance servers for various applications, including but not limited to AI training, cryptocurrency mining, and cinema-grade rendering. Clore.ai’s core strengths lie in its affordability and high cost-effectiveness. Additionally, users can earn token rewards by holding and participating in its unique “Proof of Holding” system, increasing their revenue. When users rent out their graphics card resources on the platform, the more clore tokens they hold, the more rewards they receive, without the need for locked funds and offering convenient and flexible operations.
- Render Network: Render Network (RNDR) is a distributed GPU rendering platform built using blockchain technology and ERC-20 tokens. It integrates globally unused GPU computing power to provide cost-effective cloud rendering solutions for artists and studios. The RNDR token is used to pay for rendering services and reward miners, with its core technology originating from the high-performance GPU rendering engine OctaneRender® of parent company OTOY. Facing the growing demand for GPU computing power from metaverse, AI, and VR/AR technologies, RNDR effectively reduces rendering costs and improves efficiency through a decentralized model. Currently, RNDR not only serves the traditional 3D image rendering market but also achieves innovative applications in NFT creation, AI-generated art, and immersive interactive content.Render Network connects GPU computing demanders with suppliers through dynamic pricing strategies. In December 2023, Render migrated its infrastructure from Ethereum to Solana, enhancing network performance and scalability. Additionally, Render actively explores the field of DePIN (Decentralized Physical Infrastructure Network), incentivizing individuals to participate in real-world infrastructure development through a physical proof-of-work mechanism.
- Exabits: Exabits is a future-oriented decentralized AI and high-performance computing service platform committed to building a fair, easy-to-use, and inclusive AI ecosystem. It enables users to participate in the market through Web3 identities, providing distributed GPU services, data storage, and expertise to facilitate barrier-free access to and trading of AI resources and services, including computing, storage, and models-as-a-service.
- Nosana: Nosana is a blockchain-based distributed GPU resource-sharing platform dedicated to addressing the shortage of GPUs in the market. It aggregates idle GPU resources, such as gaming PCs, mining machines, and MacBooks, to provide convenient, economical, and commitment-free GPU rental services. AI users can rent high-performance GPUs on Nosana at competitive prices, accelerating model training and usage, while GPU owners can earn additional income.
With the continuous integration of AI and blockchain technology, various innovative projects are emerging endlessly in multiple sub-fields, from GEN-AI to ON-CHAIN ML, and then to Middleware/Agent, Compute, Chain, and Data, jointly driving the wave of this technological revolution forward.
Within this thriving ecosystem, the AgentLayer project has garnered widespread attention due to its unique positioning. As the infrastructure for decentralized AI Agents, AgentLayer combines the dual characteristics of middleware and underlying public chains. It not only provides powerful support for the efficient collaboration between AI and blockchain but also offers developers vast opportunities for innovation through its flexible architecture.
Similar to AgentLayer, the MyShell project also exhibits cross-layer capabilities. It possesses rich features at the application layer while maintaining strong connectivity and scalability at the middle layer. This cross-layer design allows MyShell to achieve efficient collaboration and optimization across multiple dimensions, providing users with a smoother, safer, and more intelligent experience.
In fact, this trend of cross-layer coverage is becoming increasingly evident in current leading projects. Taking Bittensor as an example, its subnet design not only covers the application layer but also extends to the public chain layer. This comprehensive coverage enables Bittensor to innovate and optimize at multiple levels, providing users with more comprehensive and efficient services.
As technology continues to advance and application scenarios expand, we have reason to believe that more innovative and practical projects will emerge in the future, jointly promoting the deep integration and development of AI and blockchain technology. Within this process, innovative projects like AgentLayer, MyShell, and Bittensor will play a crucial leading role. They not only demonstrate the possibilities of cross-layer coverage and innovation but also drive the continuous development and improvement of the entire ecosystem through their practical applications.
The cutting-edge AI technologies and solutions exhibited at the 2024 NVIDIA GTC conference will undoubtedly inject new vitality into the Web3 world. We look forward to seeing these technologies combine with existing innovative projects to spark even more waves of excitement around AI encryption projects. Whether in generative AI, on-chain machine learning, intelligent Agents, data privacy protection, or decentralized computing networks, we will witness the tremendous potential and dynamism brought about by the integration of AI and blockchain projects.
About AgentLayer
AgentLayer, as the first decentralized AI Agent public chain, promotes Agent economy and AI asset transactions on the L2 blockchain by introducing the token $AGENT, and its AgentLink protocol supports multi-Agent information exchange and collaboration to achieve decentralized AI governance.
All Comments