AivexaNewsSearch
AI news for builders and product teamsChecked every hour

NVIDIA Announces New Planned Investments in Korean Startups at the MSIT-NVIDIA AI Startup Partnership Forum

Collected Nov 9, 2026

NVIDIA and Korea's Ministry of Science and ICT (MSIT) hosted the MSIT-NVIDIA AI Startup Partnership Forum at COEX, where NVIDIA announced an investment in Korean startup Point2 and said it plans additional investments in Motif Technologies, RLWRLD and Upstage. Ryu Je-myung, vice minister of science and ICT, joined NVIDIA leaders at the event, which brought together more than 20 Korean startups working across foundation models, physical AI, AI infrastructure and AI semiconductors. The financial terms of the Point2 investment and the size or timing of the planned deals were not stated.

The four named companies span different layers of the AI stack, which appears to be the organizing logic behind the selection. Point2 sits at the interconnect layer; Motif Technologies at foundation models; RLWRLD at robotics; and Upstage at enterprise models and document tooling.

Point2 Technology is described as an AI interconnect company building scalable, radio frequency-based interconnects that move data between chips and racks in AI factories. NVIDIA frames this as an alternative to conventional copper and optical interconnect approaches for scaling up AI clusters, aimed at the reach, performance and efficiency needed for next-generation infrastructure. In practice, interconnect is where large training and inference clusters often hit their limits: as accelerators are grouped into bigger domains, the links between them constrain how much bandwidth is available and how far a cluster can physically extend. Copper is cheap and low-latency over short distances but degrades with length; optics reach further but add cost and power. An RF-based approach is a less common middle path, and whether it delivers on efficiency claims at production scale is the open engineering question this investment is presumably betting on.

Motif Technologies is a Seoul-based AI lab developing foundation models for building AI agents that carry out complex tasks. Its Motif-3 large language model was released with open weights, meaning developers can download and run the model themselves rather than only accessing it through an API. Open weights matter for enterprises with data-residency or customization needs, and for a domestic Korean ecosystem looking to build on models it can inspect and fine-tune.

RLWRLD describes itself as a physical AI startup with offices in San Francisco, Seoul and Tokyo. It develops robotics foundation models intended to let humanoid and industrial robots handle tasks requiring five-finger dexterity, such as inserting components precisely, threading nuts onto bolts and kitting parts from moving conveyors. These are manipulation problems that classical robot programming handles awkwardly because they demand fine force control and adaptation to variation; learned models are the current bet for generalizing across them.

Upstage is a Korean enterprise AI company behind the Solar language models and document AI tools. Its products help businesses extract information from complex documents, streamline workflows and build AI applications for everyday operations. Document processing is one of the more durable enterprise AI use cases, since forms, contracts and scans are messy, high-volume and expensive to handle manually.

NVIDIA said the planned investments build on its existing work with Korea's startup and venture community, including more than 1,000 Korea-based members of NVIDIA Inception. Inception is NVIDIA's startup program offering preferred pricing, technical expertise and opportunities for investor introductions. That program is a pipeline as much as a perk: it gives NVIDIA early visibility into companies building on its platform, and the forum format suggests a structured way to convert that visibility into equity stakes and ecosystem alignment.

A likely trade-off for the startups is dependency. Taking strategic investment from a chip vendor can bring hardware access, engineering support and credibility, but it can also tie a company's roadmap to one vendor's stack at a moment when customers increasingly want portability across accelerators. For Point2 specifically, an NVIDIA-backed interconnect play is strategically interesting, since interconnect choices shape how clusters are architected around NVIDIA's own systems.

Why it matters: Developers and product teams building in Korea or selling into it get a clearer signal of which local layers NVIDIA considers strategic — interconnect, open-weight models, robotics and document AI. Expect these companies to gain easier access to NVIDIA hardware and engineering help, which in my reading could accelerate their roadmaps and make their tooling more prominent in NVIDIA's ecosystem. Teams evaluating Korean suppliers should watch whether Motif-3 and Upstage's Solar models get tighter integration with NVIDIA software, and whether RLWRLD's manipulation models show up in industrial pilots.

Read at NVIDIA Blog

Based on reporting from the original publisher. Visit the source for full context and later updates.

Publisher excerpt

Korea’s Ministry of Science and ICT and NVIDIA hosted the MSIT-NVIDIA AI Startup Partnership Forum at COEX. Representatives from the Ministry of Science and ICT and NVIDIA met with over 20 Korean startups across foundation models, physical AI, AI infrastructure and AI semiconductors. NVIDIA also announced an investment in Korean startup Point2, with additional investments […]