We are thrilled to announce the addition of Peking University (also known as Beijing University), ranked Top 10 in the world in computer science alongside MIT and Stanford, as an EdgeCloud AI customer. Theta has achieved significant momentum in both US and Korean academia, with institutions like the University of Oregon, Korea University, Seoul Women’s University (SWU), KAIST, Hankuk University, and Yonsei University utilizing its hybrid cloud GPU infrastructure to advance AI research. EdgeCloud also recently added key corporate customers including Liner, a global leader in generative AI-powered search solutions and Jamcoding, a prominent e-learning coding platform.

This marks an important milestone in Theta’s mission to support advanced AI and blockchain research through hybrid cloud GPU infrastructure. Founded in 1898, Peking University is one of China’s oldest and most prestigious research institutions, renowned for its top-tier programs in science, technology, and innovation and consistently ranked among the world’s leading universities. Known for its pioneering work in AI, distributed computing, and big data research, Peking University will leverage Theta’s high-performance EdgeCloud platform to accelerate innovation across these fields.

About Professor Zhen Xiao and His Research

Professor Zhen Xiao, a renowned researcher in distributed systems and AI, is a tenured professor at the School of Computer Science at Peking University. With a Ph.D. from Cornell University, Professor Xiao leads a lab that has published over 100 high-impact papers featured at leading conferences and journals such as USENIX ATC, SOCC, and IEEE TPDS, accumulating close to 8000 citations — a testament to his influence in research fields including distributed systems, deep learning, blockchain, and cloud computing.

The collaboration between Theta Labs and Professor Xiao started in 2022 when he joined the Theta Advisory Board and was instrumental in the early development of the EdgeCloud platform. The collaboration has also resulted in several joint research papers published in top AI and distributed systems conferences, including:

• CVPR (Conference on Computer Vision and Pattern Recognition): A prestigious forum for computer vision, focusing on AI-driven image and video analysis (paper summary).
• KDD (Knowledge Discovery and Data Mining): The premier event for data mining and machine learning research (paper summary).
• WWW (Web Conference): An influential conference on the future directions of the World Wide Web, including foundational technologies like AI and blockchain (paper summary).
• MM (ACM Multimedia): A key conference for multimedia computing, addressing AI applications across video, image, and audio data (paper summary).
• SRDS (International Symposium on Reliable Distributed Systems): A well-regarded event for research on distributed systems with an emphasis on reliability and security (paper summary).

Through these publications, Prof Xiao’s team together with Theta Labs have contributed significantly to advancements in the cross-section of AI and blockchain, specifically in adaptive defense mechanisms for AI models, distributed AI workload scheduling and placement, scalable blockchain frameworks, and cross-shard transaction optimization.

Enhancing AI Research with Theta EdgeCloud

With the support of EdgeCloud’s hybrid cloud GPU infrastructure, Peking University researchers will be able to scale their AI and distributed computing projects to new heights. EdgeCloud will enable Professor Xiao’s team to carry out state-of-the-art research initiatives in distributed machine learning, blockchain state placement, and neural network optimization.

“I’ve been working with Jieyi and the Theta engineering team for more than two years, and saw the incredible innovation and development that led to the launch of EdgeCloud. It is one of the most complex hybrid GPU infrastructure systems I’ve seen in my career,” ~ said Professor Zhen Xiao. 

 

“Now, I’m thrilled that my AI lab at Peking University and dozens of my researchers and staff members can benefit from using EdgeCloud to drive large-scale distributed AI research. We’re looking forward to replacing all of our existing lab and cloud based GPU resources with EdgeCloud’s hybrid cloud edge infrastructure.”

Professor Xiao and his entire team at Peking University have been instrumental in the development of the EdgeCloud platform and providing us invaluable AI insights in the last three years. Theta now has the most-adopted decentralized GPU platform for academia in Asia, and is quickly expanding to North America and the rest of the world. We believe the next major disruption in AI will come from infrastructure companies that harness distributed GPU and computation at scale, with minimal overhead costs and leverages excess resources and capacity. Theta’s edge network and the EdgeCloud platform is ideally positioned to lead this trend.

Link

Top Blockchain Web3 Games to Explore This Month

Here aare the best play-to-earn Web3 games this month! 🎮✨ These games let you dive into immersive experiences, join beta tests, or earn crypto while playing.

November 2024 is shaping up to be another thrilling month, especially as the bull run electrifies the crypto space. Whether you’re a seasoned player or new to the blockchain gaming world, there’s something for everyone.

Ready to level up your gaming and earning? Check out this curated list of must-play Web3 games this month! 🕹️🚀 big wins and epic adventures await!

The recent announcement of Elon Musk’s xAI Colossus cluster, featuring 100,000 NVIDIA H100 GPUs and a 150MW power supply, has set a new standard for AI supercomputing. As companies like OpenAI and other tech giants aim to match or surpass this achievement, they will likely face several challenges:

• Scalability and Integration: xAI Colossus’ massive scale and custom-built infrastructure will be difficult to replicate. Companies will need to develop innovative solutions to integrate and manage such a large number of GPUs, ensuring efficient data transfer and processing.
• Power and Cooling: The 150MW power requirement and massive heat generation will necessitate significant investments in power infrastructure and cooling systems. Companies may need to develop custom solutions or partner with specialized providers to meet these demands.
• NVIDIA Hardware and Software: xAI Colossus relies heavily on NVIDIA’s H100 GPUs and Spectrum-X ethernet networking platform. Companies may struggle to secure sufficient quantities of these components or develop compatible software and firmware to optimize performance.
• Data Center Construction and Logistics: Building and equipping a data center of similar scale and complexity will require significant resources, including land acquisition, construction, and infrastructure development. Companies may need to navigate local regulations and environmental concerns.
• Talent Acquisition and Retention: The development and operation of such a massive AI supercomputer will require a large team of experts in AI, computer architecture, and data center operations. Companies may face challenges in attracting and retaining top talent in this highly competitive field.
• Cost and Budgeting: The estimated cost of building and maintaining an AI supercomputer of this scale is likely to be substantial. Companies will need to allocate significant budgets and manage costs effectively to ensure the project’s viability.

To overcome these challenges, companies will need to invest in research and development, partner with specialized providers, and develop innovative solutions to integrate and manage their AI supercomputing infrastructure. The pursuit of xAI Colossus’ benchmark will drive innovation and advancement in the field of AI, but it will also require significant resources and strategic planning.

The Road Ahead for AI Supercomputing

The unveiling of xAI Colossus signals a transformative era for AI supercomputing, setting an ambitious benchmark that will inevitably drive the industry forward. However, the road to achieving comparable feats is fraught with challenges that demand unprecedented levels of innovation, collaboration, and strategic foresight.

Emerging Strategies to Tackle Challenges

1. Innovative Infrastructure Design
   Companies must rethink data center architectures to efficiently house and manage GPU clusters at scale. Modular and liquid-cooled designs are likely to become standard, offering improved efficiency and scalability while reducing environmental impact.

2. Alternative Hardware Solutions
   With demand for NVIDIA’s H100 GPUs skyrocketing, organizations may explore alternatives, such as AMD’s MI300 GPUs or custom AI accelerators like Google’s TPUs. Developing proprietary hardware optimized for specific workloads could also mitigate supply chain constraints.

3. Green Energy Integration
   The 150MW power consumption highlights the importance of sustainable energy solutions. Collaborations with renewable energy providers, on-site solar or wind power generation, and advanced energy storage systems will be key to balancing operational needs with environmental responsibility.

4. Decentralized AI Supercomputing
   Companies might consider distributed or decentralized approaches to supercomputing. By leveraging cloud-based GPU resources such as The Theta Network's EdgeCloud Services for computation, organizations can circumvent some of the logistical and infrastructure challenges of centralized clusters.

5. AI-Powered Operations Management
   Ironically, AI itself could assist in managing these massive systems. Intelligent algorithms can optimize GPU workloads, predict maintenance needs, and streamline resource allocation, reducing both costs and complexity.

Impact on the AI Ecosystem

As tech giants and startups alike chase the supercomputing crown, the ripple effects will reshape the AI landscape. Expect breakthroughs in fields such as natural language processing, generative AI, and robotics as these machines unlock new levels of computational power. Additionally, the competitive pressure will accelerate advancements in energy-efficient hardware, cooling systems, and AI algorithms.

Final Thoughts

Elon Musk’s xAI Colossus has redefined the possibilities for AI infrastructure, but it’s more than just a showcase of technical prowess—it’s a catalyst for innovation. The challenges it presents to competitors will inspire solutions that benefit the entire industry. As companies gear up to meet or surpass this standard, the future of AI supercomputing looks brighter than ever.

The age of AI colossi is here. Will the next giant leap come from Musk’s peers, or an unexpected disruptor? Only time will tell.