[Interview] 'Nvidia Partner' Nota to Accelerate 'AI Democratization'

Nvidia, ARM, and Intel: Major Big Tech Partners
K-Startup DeepX Joins Hands for Global Expansion
Chae Myungsu: "Contributing to AI Democratization with Optimization Technology"

Even as artificial intelligence (AI) technology advances, it is difficult for it to become widespread if hardware performance cannot keep up. While large language models (LLMs) like ChatGPT perform computations in the cloud, the majority of AI models run in device environments. There is a Korean startup that has captivated global semiconductor big tech companies such as Nvidia, ARM, and Intel with its technology that bridges this gap. This is the story of Nota.

On May 20, at the Nangang Exhibition Center in Taipei, Taiwan, where the 'Computex 2025' exhibition was being held, Nota CEO Chae Myungsu told Asia Economy, "Nota is moving toward the long-term goal of 'AI democratization,'" adding, "Computex, where the heads of chip manufacturers from around the world gather, is a place of opportunity for us."

Kim Nokwon, CEO of DeepX, and Chae Myungsu, CEO of Nota, are taking a commemorative photo after an interview with Asia Economy at the 'Computex 2025' exhibition held on the 20th at the Nangang Exhibition Center in Taipei, Taiwan. Photo by Jang Heejun

Nota achieves AI model compression and optimization through its 'NetsPresso' platform. Compression is a technology that reduces the 'size' of high-performance AI models so they can run on a variety of devices, while optimization is a software technology that enables 'efficient computation' on chips tailored to rapidly advancing AI models. While it takes years for next-generation semiconductor technology to be mass-produced, AI models evolve even faster, causing a gap. Nota's technology is specialized in bridging this very gap.

On this day, Nota signed a memorandum of understanding (MOU) with DeepX, another Korean startup that is gaining recognition for its technology in the global chip market. CEO Chae explained, "Nota excels on the software side, but we cannot get involved in the chip segment, which requires outstanding performance," adding, "Although we have worked with many semiconductor companies, DeepX is the partner that aligns with our ultimate goal of 'democratization.'"

Nota has formed partnerships with major companies such as Nvidia, Qualcomm, Hailo, and NXP, providing software technology for their chips. This means they have experience with all the leading semiconductors dominating the global market. The significance of two startups, each rising as a powerhouse on the global stage with their own technology, joining forces is therefore considerable. Although their fields and roles differ, a strong 'domestic partnership' similar to the Nvidia-TSMC ecosystem is anticipated.

CEO Chae emphasized, "Chip performance and software technology must align to create meaningful business in the market." He continued, "Many companies that receive chips and produce modules or products are based in Taiwan, and here as well, everyone asks about 'DeepX,'" adding, "This MOU will be an inflection point that will further accelerate 'AI democratization' through DeepX's outstanding chip design and Nota's optimization technology."

The DeepX booth set up at the Computex 2025 exhibition hall held from the 20th to the 23rd at the Nangang Exhibition Center in Taipei, Taiwan. On the 20th, Nota signed a memorandum of understanding (MOU) with DeepX here and agreed to support software technology for artificial intelligence (AI) optimization. Photo by Heejoon Jang

During Computex, CEO Chae plans to hold meetings with major companies such as Qualcomm, NXP, and DFI. He said, "Through collaboration with companies that will continue to grow, we aim to contribute our software technology to numerous fields including mobile, automotive, and AI PCs," adding, "In the future, when people ask, 'Which companies helped make AI technology so widespread in our daily lives?' I hope Nota comes to mind."

Nota, founded in 2015, specializes in AI model optimization and compression. The company enables efficient computation for AI model chips, making it possible to achieve high performance even on edge devices. CEO Chae Myungsu, born in 1989, earned a master's degree in Industrial and Systems Engineering from KAIST and served as a researcher at the KAIST AI Research Institute. Since joining Nota in 2018, he has been leading the company overall.