TEN Inc. Unveils Beta of RAX, the World’s First AI Infrastructure Auto-Design Service for Enterprises

South Korea-based TEN Inc. has launched the beta version of RAX, the world’s first AI infrastructure auto-design service for enterprises. The platform aims to tackle the high costs and complexity that organizations face when building AI infrastructure in today’s rapidly growing AI industry.   Over the past five years, TEN has operated AI Pub, an AI infrastructure optimization platform, providing solutions to numerous global enterprises. Drawing on the operational experience and empirical data accumulated through AI Pub, TEN identified another pressing challenge for organizations: the complexity and uncertainty of AI infrastructure design. To address this, the company developed RAX, a first-of-its-kind service that automates infrastructure design for enterprise AI workloads.   Unlike conventional solutions that rely on theoretical models or simulations, RAX automatically determines and validates the best combination of CPU, GPU, memory, storage, and network resources across an entire cluster using real AI workload data. While most existing tools offer only partial, theory-based recommendations, RAX provides a comprehensive, cluster-level design. It does so by running real workloads on TEN’s proprietary testbed, which includes a variety of hardware and software configurations, delivering highly accurate and reliable designs based on empirical results.   A TEN spokesperson said, “The success of AI services ultimately starts with the infrastructure. RAX helps companies secure the most optimized infrastructure for their AI models, eliminating uncertainty and reducing waste.” They added, “Through this, TEN has strengthened its unique, end-to-end capabilities, covering not only operational optimization but also the critical early stages of infrastructure design.”   Moving forward, TEN plans to continue driving innovation in AI infrastructure efficiency and scalability, helping more enterprises build stable and high-performing AI environments. By combining operational optimization with design services, the company aims to offer an expanded, integrated solution and extend its reach to global markets, marking the start of a full-scale international push.    정현식 기자infrastructure enterprises ai infrastructure kind service recommendations rax

2025.09.15. 1:03

Theta Network Onboards Hypernology as 40th AI Customer Powering Smart Factories and Urban Infrastructure

Theta Labs, the leader in decentralized cloud infrastructure for AI, media, and entertainment, today announced a partnership with Hypernology Inc., a fast-growing AI company driving intelligent automation across manufacturing, construction, and public safety. As part of the collaboration, Hypernology will become a valued customer of Theta EdgeCloud, leveraging its decentralized GPU infrastructure to scale their AI-powered visual inspection and safety solutions across industries.   Based in Suwon, South Korea, Hypernology has earned national recognition for its work in computer vision and AI convergence—receiving a Ministerial Award on the 58th Science Day and being spotlighted at AI EXPO KOREA 2025 for its contributions to Korea’s AI transformation.   The company’s flagship platform, hyperQ Vision, powers AI-based quality inspection, structural crack detection, and safety risk assessment—deployed widely in smart factories and urban infrastructure. By integrating Theta EdgeCloud, Hypernology now leverages high-performance decentralized GPUs to accelerate model training, reduce compute costs, and scale deployments efficiently.   "AI should be accessible, efficient, and scalable across industries—not locked behind centralized compute bottlenecks,” said Namhui Lee, CEO of Hypernology Inc. “With EdgeCloud Hybrid, we’re tapping into a decentralized infrastructure that enables us to serve more clients, train more complex models, and deploy visual AI systems faster and cheaper than ever before.”   With this partnership, Hypernology joins a global network of forward-looking institutions and organizations—including Stanford University, Seoul National University, KAIST, NTU Singapore, and enterprise and sports partners such as the NBA’s Houston Rockets - all leveraging Theta EdgeCloud to advance the frontiers of AI research and deployment.   This collaboration underscores the transformative power of combining cutting-edge AI innovation with decentralized compute infrastructure. By tapping into EdgeCloud Hybrid’s scalable GPU network, AI companies like Hypernology can accelerate development cycles, reduce operational costs, and bring AI out of the lab and into real-world applications—at a scale and speed previously unimaginable.   “Hypernology represents the next wave of AI companies bringing real-world impact through computer vision and automation,” said Mitch Liu, CEO and Co-Founder of Theta Labs. “Last week’s launch of EdgeCloud Hybrid fundamentally transforms GPU demand-supply by tapping into a global network of community run GPU infrastructure. Our dynamic GPU marketplace enables Hypernology to scale AI tools that keep workplaces safer, cities smarter, and factories more efficient.”    정현식 기자infrastructure factories gpu infrastructure cloud infrastructure theta labs

2025.07.03. 6:00