AI Model Open-Source Competition: Krafton’s ‘Raon’ Unveiling and Analysis of Market Landscape Shifts

~13 min read

AI Model Open-Source Competition: Krafton’s ‘Raon’ Unveiling and Analysis of Market Landscape Shifts

Intensifying Open-Source AI Model Competition: The Emergence of Krafton’s ‘Raon’

The artificial intelligence (AI) technology market has recently shown a clear shift from closed models to open-source models. In the past, only companies with substantial capital and human resources led AI model development, but now, open-source models, easily accessible and usable by anyone, have become a crucial driving force for AI technological innovation. Amidst this trend, the launch of ‘Raon,’ an AI model brand developed by South Korean game developer Krafton, and the open-source release of its key models, is creating significant ripples in the AI technology market. By publishing various AI models—including voice-supported LLMs (Large Language Models), real-time voice conversation models, TTS (Text-to-Speech) models, and vision encoders—on platforms like Hugging Face, Krafton is injecting new vitality into the AI technology ecosystem. Krafton’s bold decision is interpreted as a strategic move aimed not just at sharing technology but also at strengthening AI technological competitiveness and expanding the ecosystem.

The gaming industry is a sector with immense potential for innovation through AI technology. For instance, AI can be used to learn game character behavior patterns to create more natural and intelligent movements, or to automatically generate game environments, thereby saving development costs and time. Furthermore, AI can adjust game difficulty according to a user’s play style or provide personalized game content. Through ‘Raon,’ Krafton is demonstrating its commitment to leading AI technology not only in game development but also across various industrial sectors. This is interpreted as a move to accelerate AI technological innovation through openness and collaboration, moving away from closed AI model development. Indeed, Krafton is conceptualizing various projects based on ‘Raon’ and exploring new possibilities for AI technology through collaboration with external developers. These efforts are expected to have a positive impact on the entire AI technology ecosystem.

While the domestic AI technology market is still in its early stages, it has high potential for rapid growth through innovative initiatives like Krafton’s ‘Raon’ unveiling. The government is also actively implementing support policies to strengthen AI technological competitiveness, which is expected to further accelerate the growth of the AI technology market. Open-source AI models, in particular, are anticipated to play a crucial role in broadening the base of AI technological innovation by enabling various entities, including small and medium-sized enterprises (SMEs), startups, and research institutions, to easily access and utilize AI technology. Krafton’s ‘Raon’ release is projected to further accelerate this trend and contribute to enriching the domestic AI technology ecosystem.

Specific Implications and Ripple Effects of ‘Raon’s’ Open-Source Release

The open-source release of ‘Raon’ is expected to have various positive impacts on the AI technology ecosystem. First, it expands AI technology accessibility. Through open-source availability, anyone—including SMEs, startups, and research institutions—can easily utilize and research the ‘Raon’ AI models. This provides an opportunity for companies lacking large-scale capital and human resources to develop innovative services and strengthen their competitiveness using AI technology. For example, a small startup could use ‘Raon’s’ speech recognition model to develop a new voice-based service, or a research institution could leverage ‘Raon’s’ vision encoder to research image analysis technology. Thus, ‘Raon’s’ open-source release is expected to broaden the base of AI technological innovation and promote the use of AI technology across diverse fields.

Second, it accelerates AI technological innovation. A foundation has been laid for various developers to create new AI models and application services based on ‘Raon.’ ‘Raon’ is not merely a released AI model but can function as an open-source platform where diverse developers participate and collectively advance it. Developers can create their own AI models based on ‘Raon’ or contribute to improving and expanding its functionalities. This collaborative development environment will accelerate AI technological innovation and enable the use of AI technology in even more diverse fields. For instance, a developer in the medical sector could utilize ‘Raon’s’ image analysis model to develop a medical image analysis system, or an education sector developer could use ‘Raon’s’ natural language processing model to provide personalized learning content.

Third, it expands the AI ecosystem. Through ‘Raon,’ Krafton can contribute to fostering the AI technology ecosystem and strengthen its own AI technological competitiveness. ‘Raon’ can play a crucial role not just in sharing Krafton’s technology but also in building and expanding the AI technology ecosystem. Krafton can collaborate with various developers, share AI technology-related information, and offer AI technology training programs, among other efforts to vitalize the AI technology ecosystem. Furthermore, the accumulated technological expertise and experience through ‘Raon’ will contribute to strengthening Krafton’s own AI technological competitiveness. For example, Krafton could apply new AI models developed through ‘Raon’ to its game development or launch new AI-based services based on ‘Raon.’ Thus, ‘Raon’ is projected to play a significant role in strengthening Krafton’s AI technological competitiveness and expanding the AI technology ecosystem.

SK AX Launches ‘XGenicWire NPO’: Agentic AI Infrastructure Innovation

SK AX’s newly launched ‘XGenicWire NPO’ is an innovative service that automates infrastructure operations based on Agentic AI. While traditional infrastructure management largely involved manual monitoring and human intervention in case of issues, ‘XGenicWire NPO’ automates infrastructure operations by enabling AI agents to autonomously detect, analyze, and resolve problems. This is generating significant anticipation as it minimizes the potential for human error and reduces system downtime, thereby maximizing corporate operational efficiency. Notably, ‘XGenicWire NPO’ is designed to integrate with existing infrastructure management systems, allowing companies to establish AI-driven automated infrastructure management without extensive changes to their current systems.

Agentic AI is a technology that solves complex problems by utilizing AI agents that make autonomous decisions and take action. ‘XGenicWire NPO’ is designed to leverage this Agentic AI technology to automate infrastructure operations and respond swiftly to issues. For example, if a server becomes overloaded or a network problem occurs, ‘XGenicWire NPO’s’ AI agents can automatically analyze the root cause and take necessary actions to prevent system failures or facilitate rapid recovery. Furthermore, ‘XGenicWire NPO’ is designed to allow AI agents to improve their performance through learning, enabling the establishment of an increasingly efficient infrastructure management system over time.

‘XGenicWire NPO’ can be utilized across various sectors, including manufacturing, finance, and public services. In manufacturing, it can minimize potential risks through real-time anomaly detection across equipment and processes. For instance, it can monitor the temperature or vibration of production facilities in real time to detect anomalies and prevent equipment failures proactively. In the financial sector, it can implement an uninterrupted IT operating system during transaction and authentication processes. For example, if a financial transaction system experiences a failure, ‘XGenicWire NPO’ can automatically analyze the cause and switch to a backup system to minimize transaction interruptions. In the public sector, it can contribute to enhancing the stability of public services and strengthening disaster response. For example, if a government website experiences a surge in traffic, ‘XGenicWire NPO’ can automatically scale up servers to prevent service disruptions. This is a prime example demonstrating that AI technology is not limited to specific fields but can drive innovation across entire industries.

Key Features and Implementation Effects of ‘XGenicWire NPO’

‘XGenicWire NPO’ possesses various distinguishing features compared to existing infrastructure management systems. First, automated problem resolution. AI agents autonomously detect, analyze, and resolve problem situations, minimizing the potential for human error. This allows for much faster and more accurate problem-solving compared to manual monitoring and response, minimizing system downtime. For example, when network traffic surges, ‘XGenicWire NPO’s’ AI agents can automatically distribute traffic and alleviate overload, preventing system failures. Furthermore, AI agents learn from problem-solving processes and continuously improve their performance to respond more efficiently to similar issues.

Second, support for diverse operational models. Companies can choose from various operational models according to their requirements, including on-premise installation, BPO (Business Process Outsourcing), or ITO (IT Outsourcing). This enhances the flexibility of ‘XGenicWire NPO’ implementation by allowing companies to select the optimal operational model based on their IT environment and budget. For instance, companies with in-house IT personnel can opt for an on-premise model to build and operate the system themselves, while those lacking IT staff can choose BPO or ITO to operate the system with the assistance of external specialized providers.

Third, the provision of an AI Studio and GPUaaS (GPU as a Service). This offers an environment for designing and operating Agentic AI independently, along with GPU resources. This contributes to strengthening AI technological competitiveness by enabling companies to not just use ‘XGenicWire NPO’ but also to develop and utilize Agentic AI themselves. For example, companies can leverage the AI Studio to develop AI agents tailored to their IT environment and use GPUaaS to secure the computing resources needed for AI agent training and inference. This support helps companies internalize AI technology and continuously drive AI-based innovation. ‘XGenicWire NPO’ is expected to maximize the operational efficiency of corporate IT infrastructure and enhance AI technological competitiveness through these features.

Intensifying US-China AI Semiconductor Competition: Nvidia’s Crisis and Huawei’s Rise

US semiconductor export restrictions on China have brought significant changes to the Chinese AI chip market. In the past, Nvidia dominated the Chinese AI chip market with its overwhelming technological prowess and market share. However, with the US export restrictions limiting the supply of Nvidia’s high-performance AI chips, opportunities for Chinese companies have expanded. Huawei, in particular, is rapidly increasing its market share in the domestic AI chip market, emerging as a formidable competitor to Nvidia. This can be interpreted as a crucial indicator of technological hegemony competition between nations, extending beyond mere corporate rivalry. US export restrictions are accelerating the self-reliance of the Chinese AI chip market, and the AI semiconductor market is highly likely to be reshaped into a fierce competitive landscape between the US and China in the future.

The Chinese government’s active policy support and investment are also accelerating the growth of China’s AI semiconductor industry. The Chinese government has designated AI technology as a national strategic industry and is investing massive funds to foster the AI semiconductor industry. Furthermore, Chinese companies, backed by government support, are intensifying their technological development efforts, focusing on creating high-performance AI chips that can replace Nvidia’s. For example, Huawei is achieving performance comparable to Nvidia’s ‘A100’ through its self-developed ‘Ascend’ series of AI chips, which are being utilized in various fields. If this trend continues, the technological gap between the US and China in the AI semiconductor market is expected to narrow, and competition will intensify further.

Changes in the competitive landscape of the AI semiconductor market could also impact the domestic AI industry. Domestic companies must closely monitor these changes in the AI semiconductor market and strengthen their in-house AI chip development capabilities. Furthermore, amidst the technological hegemony competition between the US and China, a balanced diplomatic strategy is needed to expand AI technology cooperation and enhance AI technological competitiveness. For example, domestic companies could collaborate with US and Chinese AI semiconductor firms to jointly develop AI chip technology or share AI technology-related information. Additionally, the government should strengthen policy support for fostering the AI semiconductor industry and expand educational programs for training AI technology professionals. Through these efforts, the domestic AI industry will be able to actively respond to changes in the AI semiconductor market and strengthen its global AI technological competitiveness.

AI Technology Advancement and the Human Brain: Concerns Over Increasing AI Dependency and Expert Advice

As AI technology advances, concerns are being raised that increasing reliance on AI could diminish human cognitive abilities. Indeed, some research findings have indicated a decrease in cognitive function and memory among AI chatbot users. This suggests that as AI replaces human thought processes, the human brain might lose its capacity for independent thinking and problem-solving. For example, individuals accustomed to easily obtaining information and solving problems through AI chatbots may experience a decline in their ability to independently search for and analyze information. Furthermore, those accustomed to the convenience provided by AI might struggle with complex problem-solving or generating creative ideas.

However, neurologists advise that AI should be utilized as a tool to complement and augment human capabilities, rather than merely as a substitute. AI excels at analyzing vast amounts of data that are difficult for humans to process and performing complex calculations. By leveraging these AI capabilities, humans can enhance their creative thinking and problem-solving skills, and explore new possibilities. For example, AI can be used in the medical field to improve disease diagnosis accuracy and shorten drug development times. Additionally, AI can be applied in education to provide personalized learning content and increase learning efficiency. Thus, when utilized as a tool to complement and expand human abilities, AI can enrich human lives.

Dr. Majid Fotuhi emphasizes that while utilizing AI, consistent efforts for brain health must also be maintained. Regular exercise, sufficient sleep, healthy eating habits, and social interaction are essential elements for maintaining brain health and remain crucial in the age of AI. Regular exercise increases cerebral blood flow and promotes brain cell growth, helping to improve cognitive function. Adequate sleep is vital for the brain to process information and store memories, and sleep deprivation can lead to cognitive decline. Healthy eating habits supply the brain with necessary nutrients and help maintain brain health. Social interaction stimulates the brain and helps activate cognitive functions. By wisely utilizing AI technology and consistently practicing efforts for brain health, we can lead creative and healthy lives even in the AI era.

Specific Methods to Keep Your Brain Sharp in the AI Era

To keep your brain sharp in the AI era, various methods must be employed, including AI-powered learning, AI-powered creation, and efforts for brain health. First, AI-powered learning. Utilize AI tutors and AI-based learning platforms to enhance learning efficiency and acquire new knowledge. AI tutors provide personalized learning content tailored to an individual’s learning level and style, and analyze learning progress to adjust study plans. AI-based learning platforms offer diverse learning materials and support interaction among learners to boost learning effectiveness. For example, using an AI tutor for foreign language learning can help with pronunciation correction and grammar study, while an AI-based learning platform can improve foreign language skills through various learning resources.

Second, AI-powered creation. Engage in creative activities and generate new ideas using AI tools such as AI music composition and AI drawing. AI music composition tools automatically complete songs based on melodies or rhythms input by the user, while AI drawing tools generate new images based on text or images provided by the user. Enjoying creative activities with these AI tools can stimulate the brain and help in generating new ideas. For example, one can enhance creativity by using AI music composition tools to create their own music or by using AI drawing tools to produce unique artworks.

Third, efforts for brain health. Maintain brain health through regular exercise, sufficient sleep, healthy eating habits, and social interaction. Regular exercise increases cerebral blood flow and promotes brain cell growth, helping to improve cognitive function. Adequate sleep is vital for the brain to process information and store memories, and sleep deprivation can lead to cognitive decline. Healthy eating habits supply the brain with necessary nutrients and help maintain brain health. Social interaction stimulates the brain and helps activate cognitive functions. For example, maintaining brain health can involve exercising for at least 30 minutes daily, getting at least 7 hours of sleep per night, maintaining a balanced diet, and frequently interacting with friends and family. Through these methods, we can keep our brains sharp and lead creative and healthy lives even in the AI era.