GTC Taipei 2026: NVIDIA and NAVER Cloud’s full-stack AI partnership
NVIDIA’s annual GPU Technology Conference (GTC) came to Taipei this year, held alongside COMPUTEX, Asia’s largest IT exhibition. Anchored by a keynote from NVIDIA CEO Jensen Huang, the event brought together expert sessions on where cutting-edge technology stands today and where it’s headed—from AI factories and scaling infrastructure to agentic AI and physical AI.
Things got underway with Jensen Huang’s keynote, which laid out a blueprint for the age of the AI factory. It was here that Huang named NAVER Cloud as a key partner in global AI-native cloud. And at the NCP Summit, we shared the technology and the thinking behind how we’re building the future of the AI factory together with NVIDIA.
NAVER Cloud’s CEO on the future of the global AI factory
At the NCP Summit, NAVER Cloud CEO Kim Yuwon gave a talk titled “Scaling Sovereign AI into AI Factories,” covering the full-stack AI partnership between NAVER Cloud and NVIDIA and where the two companies are headed as they build the AI factory together. Speaking to an audience of global AI leaders, he turned heads by laying out NAVER Cloud’s full-stack capabilities across the entire AI value chain.
Here’s a quick look at the key points from the “Scaling Sovereign AI into AI Factories” talk.
Full-stack AI capabilities spanning the entire AI value chain
Our full-stack AI capabilities span the entire AI value chain. From data centers and cloud services to foundation models and AI applications, we design and operate every layer of the stack with our own technology and expertise. That edge is the product of years of steady investment, driven by a long-standing commitment to sovereign AI.
- AI services: a range of AI-powered solutions for end users
- Data: sourcing and processing high-quality AI training data
- AI backbone: developing large language models such as HyperCLOVA X
- Supercomputing infrastructure: operating a high-performance computing environment built on NVIDIA DGX SuperPOD
- Cloud infrastructure: delivering the highly scalable infrastructure of the NAVER Cloud Platform
- Data centers: building advanced hyperscale data centers such as GAK Chuncheon and GAK Sejong
Rather than chasing trends, we’ve focused on building core technical capabilities—and that philosophy still guides us today.
Where the AI journey began: The 2020 DGX SuperPOD
That journey began with a close collaboration with NVIDIA. In 2020, we became the first in the world to deploy the NVIDIA DGX SuperPOD. Working through countless rounds of trial and error together from the very start, the two companies built deep expertise in operating large-scale AI infrastructure.
Building on that experience, we’ve steadily advanced our AI infrastructure capabilities and grown them into a business. Today we provide AI models and GPUaaS to leading Korean companies and institutions—including Hyundai, Korea Hydro & Nuclear Power, and the Bank of Korea—and we’re building the National AI Computing Center, set to become core infrastructure for Korea’s AI industry.
Taking the partnership a step further
Building on this long history of collaboration and shared foundations, the two companies are now taking the partnership further. This is more than a supplier-and-customer relationship—it’s a strategic partnership to jointly develop AI technology and grow the global AI ecosystem, and it comes down to two things: deep technical integration and core business collaboration.
1) Technical integration: HyperCLOVA X and full-stack collaboration
Let’s start with the technical side. We’re advancing our own LLM, HyperCLOVA X, on the foundation of NVIDIA’s Nemotron and NeMo Framework. Across 4,224 Blackwell B200 GPUs, the two companies work closely throughout the entire model-development process—from data preprocessing to training and optimization. It’s full-stack collaboration in the truest sense, spanning the whole hardware and software stack.
We’ve also joined the Nemotron Coalition, NVIDIA’s open-model alliance. We expect the experience we gained building HyperCLOVA X, along with our expertise in running large-scale models, to make a meaningful contribution to the global open-model ecosystem and the next generation of AI.
2) The Seoul World Model for physical AI
We’re also deeply invested in world models for the era of physical AI.
“Sovereign AI models start with sovereign data.”
— Kim Yuwon, CEO of NAVER Cloud
Our Seoul World Model is built on NVIDIA Cosmos and trained on NAVER’s high-resolution map data of Seoul. Going forward, we plan to keep expanding this Cosmos-based collaboration with NVIDIA for physical AI.
From training to inference: The era of large-scale AI infrastructure
Now to the business side. In this year’s GTC keynote, Jensen Huang declared, “The inflection point for inference has arrived.” The AI industry’s center of gravity is shifting decisively from training to inference—and the numbers make the signal unmistakable. Just last year, NVIDIA projected that demand for AI computing would reach $500 billion by 2026; this year, it doubled that forecast, predicting $1 trillion by 2027. The era of large-scale AI infrastructure has truly arrived.
NAVER Cloud’s next move: Becoming Asia’s AI infrastructure hub
Our next move follows the same direction. We invested early in sovereign AI models and large-scale GPU infrastructure, and on the strength of full-stack capabilities spanning infrastructure, models, and services, we’ve established ourselves as an AI-native cloud provider that supports both general-purpose and sovereign AI. Now, by building and operating a global AI factory with NVIDIA, we aim to become Asia’s new AI infrastructure hub.
To get there, we’re expanding our business from AI-native GPUaaS into AI platforms and applications. We’ll scale our infrastructure step by step, and over the medium to long term we plan to grow into a global AI factory operator. In doing so, we’ll anchor one pillar of East Asia’s AI infrastructure—and here, NVIDIA goes beyond a technology partner to act as a key catalyst connecting global demand with NAVER Cloud.
NAVER Cloud’s strengths: Proven experience and regional partnerships
We have two clear strengths that will make this a reality.
First, proven experience across the AI stack. Drawing on what we’ve learned as an AI-native cloud provider, we can deliver solutions optimized for a wide range of enterprise workloads. Second, a track record of building sovereign AI ecosystems. Through close collaboration with regional partners, we can build AI that reflects the needs and values of each region.
NAVER Cloud’s vision: Building the foundation for the age of intelligence
The AI factory will be the core infrastructure of the future AI economy. Together with NVIDIA, we want to build that foundation and make AI more accessible, more scalable, and more valuable for customers around the world.
“Just as power plants were the foundation of the industrial age, the AI factory will be the foundation for the age of intelligence.”
— Kim Yuwon, CEO of NAVER Cloud
This is the Asian AI infrastructure hub we’re building—with our data, our infrastructure, and our technology. We hope you’ll keep following NAVER Cloud’s AI factory journey with NVIDIA.