Manta Network Partners with Aethir for Enhanced GPU Access

Theta Network has announced a significant partnership with the Liu Lab at Brandeis University, led by Professor Hongfu Liu, to utilize Theta EdgeCloud for enhancing machine learning (ML) and artificial intelligence (AI) research. This collaboration marks a pivotal moment for Theta as it solidifies its position as a leader in decentralized GPU infrastructure for academic research. The Liu Lab joins a prestigious list of institutions, including Stanford University and Seoul National University, that are leveraging EdgeCloud’s hybrid GPU capabilities to boost productivity in AI research. By integrating these decentralized resources, the lab can access scalable and high-performance computing power, which is crucial for advancing their research initiatives in data-centric learning and clustering analysis. Professor Liu emphasized the benefits of this integration, stating that the flexibility and cost-effectiveness of Theta EdgeCloud allow their team to focus on innovative research projects without the burden of managing extensive computational resources. The lab's research primarily revolves around data-centric learning, which prioritizes the quality and diversity of training data over mere algorithm refinement. This approach is essential for developing reliable and fair machine learning models, as it ensures that the datasets used are well-annotated and representative of real-world scenarios. The Liu Lab's ongoing studies in this domain include various applications such as noisy label correction and active learning, showcasing the breadth of their research capabilities. Theta EdgeCloud’s decentralized infrastructure not only empowers the Liu Lab but also supports a wider academic community by providing on-demand, high-performance computing resources. This initiative allows researchers to dynamically allocate resources, optimizing both performance and cost for large-scale projects. As Theta Network continues to partner with leading institutions, it aims to facilitate groundbreaking research in AI and machine learning, ensuring that researchers can focus on their work without the constraints of traditional computing limitations. This collaboration represents a significant step forward in making advanced AI research more accessible and efficient for academic institutions worldwide.

The Theta ecosystem is witnessing significant growth and adoption across various sectors, including sports, media, academia, and AI startups. In the latest April roundup, Theta has made headlines with its collaboration with the NBA's Houston Rockets, launching an innovative AI-powered mascot named "ClutchBot." This initiative marks a pivotal moment for Theta as it expands its influence in professional sports, showcasing the potential of AI technology in enhancing fan engagement and experience. In addition to its partnership with the Houston Rockets, Theta is also making strides in Major League Soccer (MLS) by assisting the San Jose Earthquakes in launching the league's first interactive AI agent chatbot. This development highlights Theta's commitment to integrating AI solutions into sports, further establishing Theta EdgeCloud as a leading platform for AI applications in both professional sports and esports. The platform continues to attract a growing roster of team partners, solidifying its position in the market. Moreover, academic institutions are increasingly recognizing the value of Theta's technology. Stanford University's AI Lab, led by Professor Vitercik, has begun utilizing Theta EdgeCloud for AI research, demonstrating the platform's versatility beyond sports. Theta has also introduced the first decentralized On-demand AI Model API Service, which adds more AI models to EdgeCloud. This expansion is complemented by top esports teams like NRG and EvilGeniuses launching their own Agentic AI chatbots powered by EdgeCloud. As the Theta World Tour continues at events like Paris Blockchain Week and Token2049 in Dubai, the platform is gaining more press coverage and visibility in the blockchain space.

On April 24, 2025, LBank Exchange announced the listing of Phoenix (PXT), a new digital asset designed to bridge the gap between Artificial Intelligence (AI) and decentralized physical infrastructure networks (DePIN). The PXT/USDT trading pair is now available, marking a significant step in the evolution of Web3. Phoenix aims to provide a modular and scalable framework for developing AI agents and intelligent applications directly on the blockchain, enhancing user interaction with digital assets and decentralized governance. Phoenix's ecosystem is built on the principles of composability and accessibility, allowing developers and users to create autonomous agents and decentralized applications with minimal technical barriers. The integration of AI with DePIN transforms decentralized networks into programmable economies, enabling AI agents to engage in value creation and resource coordination autonomously. This innovative approach addresses critical limitations in the current Web3 landscape, such as cross-chain interoperability and the complexity of onboarding non-technical users, while enhancing the real-world utility of decentralized systems. The native utility token, PXT, plays a vital role in the Phoenix ecosystem, facilitating economic interactions and incentivizing participation among developers, agents, and users. With a total supply of 21 billion PXT tokens, the distribution includes community incentives, ecological development funds, and allocations for team support and strategic partners. This tokenomics model not only fuels transactions and dApp operations but also fosters a sustainable DePIN economy, allowing for staking and cross-chain economic activities. Phoenix is poised to drive the next wave of adoption at the intersection of AI, DeFi, and digital infrastructure.

In a significant development for the decentralized AI landscape, io.net, a DePIN for GPU compute, has announced a strategic partnership with Mira Network. This collaboration is set to tackle the pressing challenges of AI accuracy and reliability, aiming to create scalable and decentralized solutions for advanced AI applications. By leveraging io.net's decentralized GPU infrastructure, Mira Network intends to enhance its operations while simultaneously reducing costs and latency, thus paving the way for more efficient AI implementations. A key component of this partnership is the support for Mira Network's Node Delegator Program, which allows contributors to delegate GPU resources. This initiative lowers the technical barriers for individuals and organizations eager to engage in the decentralized AI ecosystem, enabling them to support Mira's consensus operations while earning network rewards. As AI technology continues to evolve, the partnership seeks to ensure that AI can deliver consistent, reliable, and unbiased insights, a sentiment echoed by Tausif Ahmed, Chief Business Development Officer at io.net. The urgency of addressing AI's accuracy challenges is underscored by the current error rates that can reach up to 30% in tasks requiring advanced reasoning. Mira Network aims to mitigate this issue through innovative developments that evaluate AI-generated outputs across various models. Their new integrations have reportedly reduced first-pass mistakes to below 5%, with aspirations to achieve error rates of less than 0.1% through ongoing research. Stone Gettings, Head of Growth at Mira Network, emphasizes that the partnership with io.net will be crucial in further decentralizing their network and providing users with access to reliable GPU resources, thereby accelerating AI adoption in various sectors.

In the first quarter of 2025, IoTeX made significant strides with the introduction of Quicksilver, a decentralized physical infrastructure network (DePIN) module that enables AI agents to interact with real-world data. This innovative framework aims to enhance the capabilities of AI by allowing it to process real-time data and respond adaptively to its environment. Partnerships with GEODNET and Nubila have been established to integrate their devices with the ioID module, facilitating machine verifiability and verified data uploads onto the IoTeX blockchain. Despite a decline in average daily active wallets and transactions, the network still shows impressive year-over-year growth, with active wallets up 850% and transactions up 2,250% compared to the previous year. The IoTeX Network operates on a modular infrastructure that combines an EVM-compatible Layer-1 blockchain with off-chain computing capabilities. Users can stake the native token, IOTX, to vote for delegates responsible for securing the network and validating transactions. In Q1, the total fees on the network decreased by 55% quarter-over-quarter to $201,000, primarily due to significant drops in gas and decentralized exchange (DEX) fees. However, the network's staking activity remained robust, with 120 active delegates securing 3.8 billion IOTX, representing 40% of the circulating supply. IoTeX's commitment to innovation is evident in its ongoing development of the Quicksilver framework, which integrates AI with decentralized networks. The introduction of BinoAI, an autonomous AI agent, further exemplifies this effort, as it shares insights about DePIN projects on social media. Additionally, the partnerships with GEODNET and Nubila highlight IoTeX's focus on verifiability and real-world data applications, paving the way for future collaborations aimed at enhancing the DePIN ecosystem. As IoTeX continues to evolve, its emphasis on integrating AI and decentralized technologies positions it as a leader in the blockchain space, driving demand for DePIN solutions.

In a significant advancement for distributed computing, CUDOS Intercloud has introduced one-click templates that simplify the deployment of AI applications. This innovative approach focuses on accessibility and usability, allowing developers to launch applications with minimal effort. By eliminating the traditional complexities associated with AI infrastructure, such as dependencies and vendor lock-ins, CUDOS Intercloud enables instant deployment without the need for KYC or sign-ups. Users can connect their digital wallets and get started in mere seconds, which is particularly beneficial for teams needing to iterate quickly in the fast-paced AI and Web3 landscapes. The CUDOS Intercloud platform now boasts a diverse catalog of ready-to-launch applications tailored for various users, including AI developers and educators. Key offerings include JupyterLab for solo experimentation, JupyterHub for collaborative environments, and vLLM for serving large language models at scale. Additionally, tools like Ollama provide user-friendly interfaces for newcomers, while OpenManus showcases the potential of agentic AI assistants. These applications collectively form a robust foundation for a distributed AI-ready stack, enhancing accessibility and scalability for underfunded teams and global researchers. Looking ahead, CUDOS Intercloud is committed to expanding its app catalog with more open-source tools and enterprise-ready solutions. The platform aims to deepen integrations with Web3 APIs and support smart contract-based resource provisioning. As the landscape of distributed applications evolves, the one-click templates will serve as a crucial interface between users and the underlying infrastructure, driving the growth of the AI economy. CUDOS Intercloud invites users to share ideas for future templates and features, fostering a collaborative environment for innovation in distributed computing.