Snowflake Inc. is acquiring Myst AI, a startup that offers an artificial intelligence development platform for companies in the energy sector.
Snowflake announced the acquisition on Wednesday. The financial terms of the transaction were not disclosed.
Energy companies use AI models to predict electricity price fluctuations, changes in customer demand and other trends. San Francisco-based Myst offers a platform that makes it easier to develop AI models capable of making business forecasts. The startup’s platform also automates related tasks, such as the process of deploying neural networks in production.
AI models generate forecasts by analyzing historical data about a phenomenon and estimating how it may change in the future. A neural network built to forecast rainfall, for example, might analyze past precipitation measurements. The AI models used in the energy sector often use similar methods.
The historical data analyzed by AI models to generate forecasts often includes time-series records, or records that not only contain information about an event but also specify when it occurred. Myst’s platform is optimized for building neural networks that process time-series data.
According to the startup, its platform offers more than half a dozen prepackaged neural networks. Myst also provides connectors for collecting time-series data such as weather measurements. Using that time-series data, customers can train the neural networks provided by Myst to perform tasks such as predicting changes in energy demand.
The startup’s platform also helps developers test the accuracy of new AI models before deploying them. A software team using its platform can have a new neural network estimate how a past business event unfolded. The neural network’s estimate can then be compared with historical data about the event to measure its accuracy.
Myst has raised $8 million from investors since launch and says that its platform is used by multiple companies in the energy sector. According to the startup, one utility used its platform to reduce errors in some of its short-term business forecasts by 55%.
AI models are used to generate business forecasts not only in the energy sector but also other industries. Following the acquisition, Snowflake could potentially extend Myst’s platform to new use cases in additional markets.
“Time series forecasting is one of the most applied data science techniques in business,” Greg Czajkowski, the senior vice president of engineering at Snowflake, wrote in a blog post. “It is used extensively in supply chain management, inventory planning, and finance. Accurate forecasts can establish measurements to guide management, facilitate planning and goal setting, and help mitigate risk.”
Myst is the second AI startup that Snowflake has acquired in the past year. Previously, the company bought Streamlit Inc. last March in a deal reportedly valued at $800 million. Streamlit provides a popular software tool that developers use to build interfaces for AI and data science applications.