Stockmark

Stockmark is a Japanese startup specializing in natural language processing and generative AI to support corporate decision-making. Founded in 2016 by Kosuke Arima, the company is headquartered in Tokyo, Japan. The current CEO is Tatsu Hayashi. Stockmark has developed several proprietary Large Language Models (LLMs) focused on the Japanese language and business contexts, addressing the challenge that only a small fraction of generic LLM training data is in Japanese. The company has developed models with up to 100 billion parameters and has been selected for all phases of the Japanese government's GENIAC project to strengthen domestic generative AI development.

The company's offerings are centered around two main products: "Aconnect" (formerly Anews) and "Stockmark A Technology" (SAT). Aconnect is an AI-powered platform that provides business information and intelligence by aggregating and analyzing internal documents, web news, and technical literature like patents and reports. It helps organizations discover market opportunities and risks by delivering relevant information from domestic and international sources in Japanese, English, and Chinese. SAT is a Platform-as-a-Service (PaaS) that offers the underlying systems of Anews, such as information structuring, search, and classification technologies, to companies looking to implement their own generative AI solutions.

The business model for Aconnect is based on the number of users, while SAT operates on a usage-based, pay-as-you-go model determined by the volume of data processed. Stockmark serves enterprise clients, particularly in the manufacturing sector, and aims to provide an "Innovation as a Service" platform that supports the creation of new businesses. The company has raised a total of over ¥8.8 billion (approximately $58.8 million) through multiple funding rounds. Its latest funding was a Series D round in October 2024, raising ¥4.5 billion led by Polaris Capital Group, with plans to use the capital for R&D, talent acquisition, global expansion, and potential M&A activities.

Keywords: natural language processing, generative AI, large language model, business intelligence, market analysis, knowledge management, enterprise AI, text mining, data structuring, Aconnect, Anews, Stockmark A Technology, Japanese AI, corporate decision-making, innovation support, sales enablement, market trends, document analysis, AI agent, PaaS