August 17, 2023

Anthropic Partners with SKT for Telecom Language Model

Anthropic is a generative AI provider engaged in the research and development of a variety of different large language models, with an eye towards preserving AI safety. Its primary product is Claude, a chatbot and competitor to OpenAI’s ChatGPT that runs on a different framework of AI model design known as ‘constitutional AI.’

Anthropic has been securing funding over the last several quarters, with several millions coming in from various sources. With this funding behind it, Anthropic is emerging as a major player alongside the likes of OpenAI. While OpenAI is closely partnered with Microsoft for delivering its AI services, Anthropic has received considerable funding and support from Google in the past.

Generative AI and Telecom

Anthropic is working to build out the capabilities of the telecom industry by leveraging large language models. Its latest round of funding comes from South Korea Telecom (SKT), which is investing $100 million to support the development of telecommunications LLMs. SKT seeks to build a language model that is custom-configured to support the telecommunications industry and its challenges.

SKT plans to combine its own Korean language LLM with the general capabilities offered by Anthropic to create an AI application specifically designed for the telecom industry. Business leaders should expect more partnerships like this one to emerge in the rapidly developing generative AI space as specific industries with access to privileged, high-quality data, recognize the power of combining this data with general purpose base AI models to achieve a highly-specific AI application.

Bottom Line

Anthropic is an emerging leader in the generative AI market that continues to receive funding for its innovations in model design and development, and its continued work in the enterprise. This deal foreshadows the rise of more specific generative AI models, powered by enterprise data and geared towards specific verticals.

Anthropic Partners with SKT for Telecom Language Model