Matchain Partners with io.net to Enhance Decentralized AI Infrastructure

Exabits, a crypto-AI startup focusing on tokenizing GPUs, has successfully raised $15 million in a seed funding round led by Hack VC. This funding round has propelled Exabits' valuation to $150 million. The startup, co-founded by Dr. Hoansoo Lee, initiated fundraising in July and concluded the round in October. While the specific investors beyond Hack VC remain undisclosed, Lee mentioned that a token launch is on the horizon. With this seed funding, Exabits has accumulated a total of $20 million in funding, building upon previous rounds earlier this year. Exabits' primary objective is to decentralize computing by tokenizing GPUs, which it considers as 'the currency of the digital age.' Lee emphasized the exponential growth potential of GPU compute performance compared to traditional assets like bitcoin and gold. The startup's unique GPU tokenization model aims to provide direct exposure to GPU compute assets, fostering a liquid market, revenue generation, and rewards for token holders. Despite the token not being launched yet, Exabits has witnessed substantial revenue growth, with a 300% quarterly increase on average, amounting to $10 million in annual recurring revenue (ARR).

The integration of confidential decentralized computation (CDC) and GPU acceleration is transforming the landscape of secure and efficient data processing. These technologies offer a groundbreaking approach to protecting sensitive information while achieving unparalleled computational performance. By harnessing CDC to ensure data privacy and leveraging GPUs for high-speed processing, we can securely manage and analyze data across distributed networks, unlocking new possibilities for innovation. ## Confidential Decentralized Computation [Confidential decentralized computation](https://www.twentysix.cloud/blog/articles/confidential-virtual-machine-overview/), or CDC, is a paradigm shift in data processing. It enables secure computation on sensitive data without compromising its privacy. By leveraging cryptographic techniques and secure hardware enclaves, CDC ensures that data remains encrypted throughout the entire computation process, even from the computing nodes themselves. This approach mitigates the risk of data breaches and unauthorized access. The Power of GPUs, originally designed for rendering graphics, have evolved into powerful parallel processing engines. Their ability to perform numerous calculations simultaneously makes them ideal for accelerating a wide range of computational tasks, from machine learning and scientific simulations to data mining and cryptography. By integrating GPU acceleration into CDC, we can unlock the full potential of both technologies. This combination offers several significant advantages: * Enhanced Privacy: CDC ensures that data remains encrypted during computation, while GPU acceleration can be used to perform complex cryptographic operations efficiently. * Improved Performance: GPUs can significantly accelerate the execution of computationally intensive tasks, such as training large machine learning models or breaking cryptographic ciphers. * Scalability: CDC can be scaled across multiple nodes, and GPU acceleration can be used to distribute the workload among these nodes, further improving performance and scalability. * Reduced Latency: By offloading computationally intensive tasks to GPUs, we can reduce the overall latency of the computation process. ### Our approach: QEMU Hypervisor and PCI-Passthrough To realize the benefits of CDC and GPU acceleration, we employed a novel approach using the [QEMU hypervisor](https://www.qemu.org/) and PCI-passthrough technology. This approach allows us to securely isolate GPU resources within a confidential computing environment. * QEMU Hypervisor: QEMU, a versatile virtualization platform, provides a secure and isolated environment for running virtual machines. By leveraging QEMU’s capabilities, we can create a trusted execution environment (TEE) for confidential computation. * PCI-Passthrough: PCI-passthrough enables direct access to hardware devices, such as GPUs, from within a virtual machine. This allows us to harness the full power of the GPU for accelerating computations within the TEE. By combining QEMU and PCI-passthrough, we can create a powerful and flexible platform for [confidential decentralized computation](https://aleph.im/) with GPU acceleration. This approach offers a high degree of security and performance, making it suitable for a wide range of applications, including: * Secure AI and Machine Learning: Training and deploying machine learning models on sensitive data without compromising privacy. * Private Data Analytics: Analyzing large datasets while preserving data confidentiality. * Blockchain and Cryptocurrency: Securing blockchain transactions and mining operations. * Secure Remote Rendering: Rendering complex graphics and virtual reality experiences remotely without exposing sensitive data. ### Conclusion The combination of confidential decentralized computation and GPU acceleration marks a pivotal advancement in secure and efficient data processing. By capitalizing on the strengths of these technologies, we can tackle pressing challenges in data privacy, security, and computational performance. Our innovative use of QEMU and PCI-passthrough offers a flexible and scalable framework, unlocking the transformative potential of secure data computation across diverse applications.

The artificial intelligence (AI) sector is experiencing rapid growth, yet many companies still rely on centralized services like Amazon Web Services (AWS) for their cloud computing needs. This reliance can lead to high costs and potential outages. Wes Levitt, head of strategy at Theta Labs, argues that decentralized cloud computing offers a solution to these challenges. In a recent episode of The Agenda podcast, Levitt discussed how decentralization not only reduces costs but also enhances reliability and flexibility for users. He emphasized that by moving away from a few centralized points of failure, Theta Labs aims to democratize access to computing power, particularly in the AI and academic sectors. Theta Labs is the driving force behind the Theta Network, which positions itself as a decentralized cloud solution for AI, media, and entertainment. Levitt noted that while the company initially focused on media, it has seen a significant increase in AI-related clients, especially from academic institutions. He highlighted partnerships with top Korean universities and the University of Oregon, explaining that decentralized cloud services provide a cost-effective alternative for organizations that cannot afford AWS. This flexibility is crucial for academic institutions, which often need to scale their computing resources rapidly for research projects before conferences and then scale back down afterward. The decentralized AI niche has seen remarkable growth, with the market capitalization of AI and big data crypto projects skyrocketing from $16.17 billion in December 2023 to over $70 billion by December 2024. Despite this surge, Levitt clarified that the goal of Theta Network is not to eliminate AWS or other centralized cloud providers. He acknowledged that there will always be use cases for centralization, particularly for specific applications requiring focused resources. However, he remains optimistic about the potential for decentralized AI to compete with centralized solutions in the future, suggesting that as demand for decentralized services grows, they could become equally significant in the tech landscape.

The telecom industry is on the brink of a transformative era with the advent of 6G technology, which promises to redefine connectivity and innovation. The JOINER project, the UK’s first national accelerator program for 6G and beyond, aims to unite researchers, businesses, and government entities to address significant challenges in the telecom sector. By fostering collaboration and breaking down silos, JOINER creates a shared platform that facilitates resource sharing and the development of innovative solutions, ultimately driving the industry forward. A standout feature of the JOINER initiative is the implementation of Hypervisor as a Service (HaaS), based on Weaver Labs’ Cell-Stack software. This service introduces a layer of cloud automation that simplifies the management of complex Telco-Cloud platforms, such as Kubernetes and OpenStack. By enabling zero-touch deployment and management, HaaS allows researchers to concentrate on innovation without the burden of technical complexities associated with cloud infrastructure. This streamlined approach enhances collaboration and accelerates the development of new technologies, positioning the UK as a leader in telecom innovation. JOINER's mission extends beyond HaaS, as it aims to facilitate large-scale research, create shared telecom testbeds, and foster collaboration among various stakeholders. With plans to host experiments across multiple interconnected locations and deliver the UK’s first full-scale 6G demonstration by March 2025, JOINER is set to play a pivotal role in shaping the future of telecom. By training the next generation of professionals and strengthening the UK's leadership in global telecom standards, JOINER, supported by Weaver Labs, is paving the way for a smarter, more connected world powered by 6G and beyond.

In the evolving landscape of Web3, data ownership has emerged as a critical concern for developers and users alike. As decentralized applications (dApps) gain traction, questions surrounding who can access user data, how it is utilized, and whether users can reclaim control after sharing become paramount. The promise of Web3 is to empower users with control over their personal information, necessitating a shift in how developers approach data security and governance. iExec offers a suite of solutions designed to address these challenges, ensuring that users retain full control over their data even after it has been shared. Developers face several key challenges regarding data ownership, including security, control, and trust. Users often hesitate to engage with dApps due to concerns about unauthorized access to their data. iExec's tools, such as DataProtector, provide robust encryption for both data storage and sharing, allowing users to encrypt files before sharing them on decentralized platforms. Additionally, iExec's on-chain data tokenization system enables users to transform their information into digital assets, granting them the ability to grant or revoke access as needed. This ensures that users can share sensitive information, like medical records, while maintaining control over who can access it and for how long. Moreover, iExec's privacy-enhancing tools, such as Privacy Pass, empower Web3 users to monetize their personal data without sacrificing ownership or privacy. By allowing users to participate in targeted marketing campaigns while keeping their personal information confidential, iExec facilitates a secure environment for data monetization. The combination of blockchain technology and Trusted Execution Environments (TEEs) ensures that sensitive data remains protected during processing, allowing third-party applications to interact with encrypted datasets without ever accessing raw information. As developers navigate the complexities of data ownership in Web3, iExec provides the necessary tools to build trust and empower users in the decentralized ecosystem.

Decentralized Physical Infrastructure Networks (DePIN) are emerging as a transformative force in the blockchain landscape, utilizing decentralized protocols to manage physical facilities and address inefficiencies in centralized resource networks. Coined by the Web3 media and research firm Messari, DePIN has gained traction since its introduction, particularly in the latter half of 2023 and throughout 2024. These projects aim to mitigate issues such as human error, malfunctions, and financial opacity, offering more secure and cost-effective alternatives. With approximately 296 active DePIN projects currently, this sector is witnessing exponential growth, indicating a robust interest in decentralized solutions across various industries, including AI, mobility, and health. The evolution of DePIN can be traced back to early blockchain projects like IOTA and Filecoin, with Helium's launch in 2019 marking a significant milestone for decentralized connectivity. As the term DePIN gained popularity, it encompassed a growing number of projects that leverage blockchain for data management and incentive distribution. By employing a 'plug and play' system, DePIN networks facilitate the interconnection of independent infrastructures, such as IoT devices and decentralized GPU grids. This innovative approach enables efficient data processing and administrative functions, allowing for diverse applications ranging from decentralized ride-sharing to computing resource networks. Currently valued at $33 billion, the DePIN market is thriving, with several projects achieving market capitalizations exceeding $1 billion. As we look ahead to 2024 and beyond, key developments to watch include the integration of AI, the emergence of multi-token ecosystems, and the expansion into trade finance. Despite facing challenges such as legal uncertainties and the complexity of blockchain technology, DePIN networks are poised for increased adoption across various sectors. By removing intermediaries and creating balanced networks, DePIN could redefine resource management and utility in the decentralized world, making it essential for potential investors to conduct thorough research before engaging with these innovative projects.