Meta has transferred its popular open source PyTorch artificial intelligence framework to the Linux Foundation, establishing the PyTorch Foundation to "democratize access" to the technology and "accelerate AI research."

More than 18,000 companies, including Tesla, use PyTorch, one of the top AI frameworks, to power their Autopilot driverless car technology. Although Meta developed it, several tech behemoths, including Amazon, Google, and chipmaker AMD, have made sizable investments in and contributions to the project. In addition, the platform supports "tensors," which enable data manipulation in ways not possible in other frameworks and allows researchers and developers to use Python as the primary programming language for AI models.

For the typical PyTorch user, things should stay pretty much the same; PyTorch will continue to be open-source and expand within the community it has built over the years. Instead of being solely owned by Meta, PyTorch will be able to grow more quickly and achieve its full potential with new sponsorship opportunities and open collaboration between industry leaders.

Furthermore, representatives from AMD, Google, Amazon, Meta, Microsoft, and Nvidia will sit on the new foundation's governing board, but Meta hopes to expand that eventually. Since its inception, it has partnered with the larger AI community to develop its framework for AI research, making it open source and inviting other organizations to contribute. Since its development, more than 150,000 projects have been on GitHub. In addition to using PyTorch as its main AI framework, Meta says it will continue to make significant contributions to the project. Moreover, it announced last year that it would move all its AI models to the platform.

The new funding opportunities from not only its six prominent members but also from outside companies will propel PyTorch's growth even further than before. In addition, as a member of the Linux Foundation, PyTorch will have access to more community events and collaboration opportunities, as well as training, certifications, and other community dynamics. Furthermore, Meta's decision to release PyTorch from their sole control demonstrates their trust in the framework, trust in the industry to know what's best for it, and recognition that for PyTorch to truly reach its full potential, it must be democratically owned.