SK Telecom Invests $3 Million in Twelve Labs to Advance Multimodal AI for Video Analysis

SK Telecom, South Korea’s telecommunications provider, has announced a $3 million investment in Twelve Labs, an AI video analysis startup based in San Francisco. This partnership will integrate Twelve Labs’ advanced multimodal video understanding technology into SK Telecom’s AI agent services. 

Founded in 2021, Twelve Labs specializes in developing AI models capable of comprehending and searching video content in a human-like manner. The startup gained attention earlier this year when it secured $50 million in Series A funding, led by Nvidia’s NVentures and New Enterprise Associates (NEA).

Twelve Labs’ technology can interpret and match video content to textual descriptions, enabling more accurate video search and understanding. For example, if a user provides a written description of a movie scene, the AI can identify the film and pinpoint the exact moment. With this investment, SK Telecom looks to expand its AI capabilities, strengthen its multimedia services, and enhance user experiences.

Twelve Labs has developed multimodal AI models that enable human-like video comprehension, semantic search, and video analysis. Their technology powers solutions like “Marengo” for image search and “Pegasus” for video-based question-and-answer systems. The startup has been recognized globally and named one of the top 100 AI startups by CB Insights for three consecutive years.

Through this partnership, SK Telecom aims to integrate Twelve Labs’ technology into its AI Agent service, a virtual assistant designed to enhance user interactions. By combining the strengths of both companies’ AI expertise, SK Telecom expects to strengthen its AI-driven offerings and improve its services. The collaboration will also explore multimodal AI applications in security and public safety, including AI-powered surveillance systems. 

Traditional surveillance systems often require human operators to sift through hours of CCTV footage. Still, Twelve Labs’ AI models can quickly identify key incidents, actions, and individuals, offering more efficient and effective monitoring.

The collaboration extends beyond security, as both companies plan to develop technologies that apply Twelve Labs’ multimodal AI to various industries. These innovations will enhance AI Agent’s functionality and provide value in other sectors, such as enterprise-level video archives and public safety applications. Twelve Labs’ ability to perform precise semantic search, summarization, and analysis positions the startup as a powerful player in the AI market, particularly in industries requiring vast video data processing.

Additionally, Twelve Labs will join the K-AI Alliance, a Korean initiative to support AI innovation. Through this partnership with SK Telecom, Twelve Labs intends to extend its video foundation models to a wider array of industry applications, driving real-world value in daily workflows. 

Jae Lee, CEO of Twelve Labs, expressed enthusiasm about this partnership’s potential to contribute to the AI ecosystem, while SK Telecom’s Lee Jae-shin highlighted the importance of this collaboration in strengthening the company’s position in the multimodal AI sector.

Twelve Labs’ multimodal AI technology combines various data types to enhance video understanding and analysis, including images, text, audio, and video. This model analyzes video content and converts it into text, edits short-form videos, and categorizes content based on specific criteria. It offers practical applications in areas such as YouTube Shorts creation, advertising strategies, and even identifying missing persons by analyzing CCTV footage. 

To optimize its video analysis capabilities, Twelve Labs has integrated Nvidia’s advanced GPU technology, including the NVIDIA H100 Tensor Core and NVIDIA L40S GPU, into its platform, which significantly enhances the performance of its models.

The collaboration between SK Telecom and Twelve Labs aims to further advance these technologies, particularly in security and public safety. By leveraging Twelve Labs’ multimodal AI, SK Telecom plans to improve AI-powered surveillance systems, enabling faster and more accurate identification of key events, movements, and individuals from CCTV footage. This shift will enhance security operations by reducing the need for human operators to monitor extensive video feeds manually.