Krafton Inc. (CEO Chang-han Kim) unveiled its AI innovation technology CPC (Co-Playable Character, hereinafter CPC) for the first time together with NVIDIA (CEO Jensen Huang) at 'CES 2025', the world's largest IT and electronics exhibition held in Las Vegas, USA.
This collaboration was achieved by combining the strengths of Krafton, with its unique game development capabilities and differentiated AI technology, and NVIDIA, a global leader in AI semiconductors. At CES 2025, Krafton, together with NVIDIA, unveiled CPC, an innovative AI technology specifically designed for gaming, signaling the future direction of the gaming industry.
CPC is a new concept character that can interact with game users, based on the On-device Small Language Model (SLM[1] for Gaming) built with NVIDIA ACE technology. Unlike existing Non-Player Characters (NPCs), CPC converses and cooperates with users, and recognizes situations and responds flexibly like a human. Krafton plans to expand the application of CPC to various games, including the PUBG IP franchise and inZOI, and lead a transformation in the user experience.
Kim Chang-han, CEO of Krafton, said, “CES 2025 is a meaningful event that showcases AI technology jointly developed by Krafton and NVIDIA and presents the direction of the gaming industry.” He added, “We will continue to work closely with NVIDIA to expand user experience with differentiated AI-based technologies, including CPC, and open a new future for the global gaming industry.”
“AI is transforming how games are developed and played,” said Keita Iida, vice president of developer partnerships at NVIDIA. “By introducing innovative AI technologies like NVIDIA ACE into Krafton’s various live games, we will enable CPC to deliver new, unique experiences.”
Meanwhile, since establishing its Deep Learning Division in 2022, Krafton has been developing various core AI technologies, including natural language processing (LM/NLP), vision & animation (3D Vision & Animation), speech recognition and generation (STT/TTS), reinforcement learning (RL), and multimodal models. Furthermore, the company has demonstrated its research competitiveness by publishing numerous papers in leading AI conferences, including NeurIPS, ICML, and ICLR.
- See more related articles
You must be logged in to post a comment.