[Edaily Reporter Ahn Yu-ri] The three-dimensional (3D) virtual worlds once used for game development are now evolving into training grounds for robotic artificial intelligence (AI). This is why #Krafton and #NC are aggressively investing in physical AI and robotics beyond their existing gaming businesses. NVIDIA is also expanding its collaboration with these companies, viewing them not merely as GPU (graphics processing unit) customers, but as key partners to jointly lead the era of digital twins and robot AI.

According to the gaming industry on the 9th, Krafton and NCSoft have recently expanded their dedicated physical AI teams to double-digit sizes and are accelerating the recruitment of related personnel. Both companies are fostering robotics and physical AI businesses as future growth engines, primarily through their subsidiaries.

After establishing the robotics subsidiary "Ludo Robotics" in San Francisco, U.S., last year, Krafton began actively recruiting talent in earnest early this year. According to LinkedIn, Ludo Robotics has a workforce of 11 to 50 employees and is actively hiring R&D personnel by offering a compensation package on par with that of local big tech companies. Recently, the company has accelerated its organizational expansion by recruiting postdoctoral researchers in the "Sim-to-Real" field, which involves applying AI trained in virtual environments to real robots.

Krafton recently signed a memorandum of understanding (MOU) with Hanwha Aerospace to establish a joint venture (JV) for the joint development of physical AI, marking a step toward business expansion. Last year, the company established a joint venture with SOCAR to secure autonomous driving data and made a strategic investment of 65 billion won. Krafton plans to utilize the massive amounts of autonomous driving data accumulated during vehicle operations for future research and development in physical AI and robotics.

NC AI, NC’s AI subsidiary, also established the “Physical AI Lab” earlier this year and has expanded its team size by double digits. The company plans to continue strengthening its R&D capabilities through additional hiring. In particular, this year it is focusing on commercializing physical AI that can be practically applied in industrial settings, moving beyond the research phase.

Following the unveiling of the World Foundation Model (WFM)—which serves as the “brain” of robots—last March, NC AI is now actively seeking clients, particularly in the manufacturing, logistics, and defense sectors. Recently, it secured a contract from Hanwha Ocean for the “Development of a Vision-Based Dedicated Welding Model and a Collaborative Robot-Based Autonomous Welding Model,” and was also selected as the final implementing agency for a national R&D project commissioned by the Agency for Defense Development (ADD). The company is being recognized for its ability to consistently demonstrate physical AI capabilities applicable in industrial settings.

Game Companies Becoming Key Collaboration Partners in the

AI Ecosystem

Krafton and NCSoft also participated as key partners in the “Korea AI Ecosystem Reception” hosted by NVIDIA on the 8th, which invited representatives from domestic AI and robotics companies.

In the past, the relationship between NVIDIA and game companies was similar to that of a technology company supplying GPUs and gaming PCs and a client company utilizing them to develop games. In fact, since the early 2000s, NC has actively adopted NVIDIA’s GPUs and graphics technology during the development of major games such as ‘Lineage II,’ ‘Aion,’ and ‘Blade & Soul,’ thereby enhancing game quality and service stability.

However, with the dawn of the AI era, the relationship between the two sides is evolving beyond that of a simple supplier and customer into a strategic partnership. A prime example is the AI character technology “CPC (Co-Playable Character),” jointly developed by Krafton and NVIDIA. While traditional NPCs (non-player characters) were limited to repeating predetermined lines and actions, CPC is characterized by its ability to make decisions, converse, and interact actively with users.

This technology utilizes NVIDIA’s digital human platform, ‘ACE (Avatar Cloud Engine),’ and has been implemented in Krafton’s life simulation game, ‘inZOI.’ Krafton also plans to introduce ‘PUBG Ally,’ an AI companion character, in ‘PUBG: Battlegrounds’ later this year.

Krafton Chairman Jang Byung-gyu met with reporters on the 7th following a meeting with NVIDIA CEO Jensen Huang at a PC bang and stated, “NVIDIA’s ‘RTX Spark,’ set to be unveiled this fall, is a chip designed to combine gaming and AI,” adding, “Krafton has been developing PUBG Ally in line with this for the past one to two years. "We have now entered an era where games and AI are discussed together," he said.

Game Engines as Textbooks for Robot Learning

The virtual world-building technology that game companies have accumulated over decades is emerging as a core asset in the era of humanoid robots. Game engine-based three-dimensional (3D) virtual environments can not only implement physics laws similar to those in the real world but also calculate interactions between numerous objects and characters in real time. Thanks to this, robots can repeat the training that would otherwise take months to complete in actual factory or urban environments much more quickly and safely within a virtual space.

Synthetic data, which has recently garnered significant attention in the AI industry, is also closely linked to gaming technology. This is because data that is difficult or costly to obtain in the real world can be generated in large quantities within virtual environments. This is why the value of gaming technology is being reevaluated as a means of securing training data to enhance the performance of robots and autonomous driving AI.

The digital twin platform "Omniverse," led by NVIDIA, is also evolving based on these virtual environment construction technologies. Consequently, game companies are no longer merely customers purchasing GPUs; they are establishing themselves as key partners in implementing actual services within digital twin and robot simulation environments.

Jeon Seong-min, a professor in the Department of Business Administration at Gachon University, stated, “NVIDIA is seeking to secure global partners capable of implementing digital twins,” adding, “From NVIDIA’s perspective, game companies with the capability to build and operate virtual worlds are extremely important partners.”

Seeking New Growth Engines Amid a Slowing Gaming Market

Behind game companies’ focus on physical AI lies a strategy to reduce their reliance on the traditional gaming business. While the gaming industry experiences significant earnings volatility depending on the success of new titles, the robotics AI and industrial AI markets offer the prospect of a stable revenue structure centered on business-to-business (B2B) transactions.

The market is also taking note of this shift. Shares of NCSoft and Krafton surged sharply on the 4th immediately after news broke of their meeting with NVIDIA CEO Jensen Huang.

As of the closing price on the 9th, NC’s stock stood at 264,500 won, marking a 30.3% increase over the past six months. During the same period, Krafton’s stock price fell by only 0.4% to 244,500 won, showing a relatively steady trend.

In contrast, stock prices across the domestic gaming industry were sluggish. Netmarble’s stock price stood at 40,600 won on the 9th, down 17.8% over the past six months. Nexon, listed on the Tokyo Stock Exchange, also closed at 2,194 yen, down 42.4% over the same period.

Analysts suggest that as AI and semiconductors emerge as key investment themes in the market, the potential for companies to transform from simple game developers into AI firms is being reflected in their stock prices.

Yoon Seok-bin, a special professor at Sogang University’s Graduate School of AI and Software, stated, “Game companies are now playing a role within the NVIDIA-centric AI ecosystem,” adding, “There are many areas where the unique technological capabilities of the Korean gaming industry can create synergy with NVIDIA.” He continued, “Recent stock price trends should also be considered in light of the supercycle centered on AI and semiconductors, rather than the gaming industry itself.”