Seoul, South Korea – August 29, 2025 – Lablup, Inc. announced today that it will participate in AI Infra Summit 2025, showcasing its comprehensive AI infrastructure management capabilities. 

AI Infra Summit 2025 brings together over 3,500 global AI professionals and 100+ partner companies to address the full spectrum of AI infrastructure from hardware to data centers and edge environments. Lablup will demonstrate at booth #620 the beta version of its AI-native unified platform AI:DOL alongside Backend.AI running on a miniature AI data center built with palm-sized NVIDIA Jetson Orin Nano devices. 

Lablup unveils the beta version of its AI-native platform 'AI:DOL (AI Deployable Omnimedia Lab)'. AI:DOL combines Backend.AI Core with PALI-based services to deliver a generative AI development platform accessible via web browser without installation. It supports developers from beginners to experts, enabling AI model deployment in a low-code environment. Through Backend.AI Continuum, AI:DOL integrates on-premises and cloud resources, enabling streamlined development and deployment of AI applications for internal and external clients. Currently focused on chat functionality, AI:DOL plans to expand into image, video, and AI coding capabilities to become a comprehensive AI development ecosystem. 

Lablup also showcases the 'Mini AI Factory' and demonstrates how the same Backend.AI platform scales from edge devices to hyperscale data centers with over 1,500 GPUs. Visitors can witness modular scalable architecture, heterogeneous AI accelerator management, and patented container-level GPU virtualization technology. Showcase highlights the infrastructure requirements for Sovereign AI and the importance of compliant, efficient infrastructure management solutions. 

Lablup participates in a South Korean government-led AI foundation model project managed by the Ministry of Science and ICT as part of the Upstage consortium. The company operates NVIDIA B200 GPU infrastructure in SK Telecom's Haein cluster and collaborates with Upstage, NVIDIA, and SK Telecom to research stable and accelerated large-scale foundation model training technologies. This involvement supports Korea's national Sovereign AI initiative while strengthening Lablup's global competitiveness through large-scale AI training technology advancement. 

"The U.S. market is experiencing unprecedented demand for AI infrastructure and managed backend platforms that align with rapidly growing AI workloads," said Jeongkyu Shin, CEO of Lablup, Inc. "Growing focus on Sovereign AI further accelerates AI infrastructure demand. Yet, scaling AI reliably across diverse and distributed environments remains a major challenge. Backend.AI addresses this need with a globally deployable, scalable solution that enables confident AI expansion at every stage." 

Lablup aims to strengthen its global presence while highlighting its role in shaping the next generation of enterprise AI infrastructure.  

About Lablup

Lablup, Inc. develops Backend.AI, an AI infrastructure operating platform built for every scale. Continuously building a product that makes it easy to operate and scale AI. We strive to address the challenges of sharing scientific and engineering research through a standardized platform, while also automating the entire process—from preparing research to delivering final code. With its innovative technology, Lablup enables the global AI community to accelerate discovery and innovate with confidence. 

Media Contact

DevRel Team | Lablup, Inc. 
Representative Email: wyyoo@lablup.com 
Website: www.lablup.com | www.backend.ai