OpenAI Inc. is reportedly seeking to raise a funding round from Microsoft Corp., which invested $1 billion in the artificial intelligence research group three years ago.

The development was reported by The Information and the Wall Street Journal on Thursday. OpenAI is said to be in “advanced talks” with Microsoft about the funding round. It’s currently unclear how much financing the research group is seeking to raise or at what valuation.

According to The Information, OpenAI received an implied valuation of nearly $20 billion in a stock sale that took place last year. The participants in the stock sale reportedly included Sequoia Capital, Tiger Global Management, Bedrock Capital and Andreessen Horowitz.

OpenAI earlier secured a $1 billion investment from Microsoft in 2019. The following year, Microsoft detailed that it has built a supercomputer in its Azure public cloud platform to support OpenAI’s machine learning research. The supercomputer, which features 10,000 graphics cards and 285,000 central processing unit cores, is used by the research group for tasks such as training neural networks.

Microsoft’s collaboration with OpenAI has expanded to several additional areas since its $1 billion investment in 2019.

Earlier this month, Microsoft announced plans to make OpenAI’s DALL-E 2 AI system available to customers of its Azure cloud platform. DALL-E 2 can automatically generate images based on user prompts. Microsoft also plans to incorporate DALL-E 2 into a new design application called Microsoft Design as well as its Bing search engine.

Previously, Microsoft’s GitHub subsidiary used another OpenAI-developed AI system called Codex to build a productivity tool for developers. The tool, GitHub Copilot, can automatically generate snippets of code to save time for developers.

As part of its research, OpenAI develops not only AI systems but also tools for building them. The group last July debuted Triton, a specialized programming language for writing machine learning software. Triton promises to ease several of the tasks involved in optimizing a neural network’s performance.

Alongside Triton, OpenAI has developed software to ease AI training, which is among the most complicated aspects of machine learning projects. Earlier this year, OpenAI and Microsoft introduced a jointly developed technology that makes it possible to train neural networks more efficiently.

A new funding round from Microsoft could help OpenAI expand the scope of its machine learning research to more areas. Moreover, such an investment may indirectly benefit Microsoft’s own product development efforts. Microsoft has incorporated OpenAI-developed neural networks into several of its products and could take a similar approach with new AI systems that the research group releases in the future.

Google LLC, one of Microsoft Corp.’s top competitors in the cloud market, has likewise made AI research a priority. The company’s research division actively invests in the development of new machine learning technologies. Google parent Alphabet Inc.’s DeepMind subsidiary is also working to advance the capabilities of AI software.

Over the years, technologies developed by DeepMind have been applied by Google in several areas. The search giant uses DeepMind-developed neural networks to extend the battery life of Android devices and deliver app recommendations in the Play Store. Another joint project led to the development of an AI system that has helped Google increase the energy efficiency of its data centers. 

Image: OpenAI