In 2009, half of open source code was licensed under the GPLv2 license, the canonical copyleft license. Every other license had less than 10% market share. Over the years, the MIT license and other permissive licenses kept climbing at the expense of the GPLv2. As of today, the MIT license is the leading license with more than 32% market share in absolute numbers, with the GPLv2 license having fallen below 20%.

These are absolute numbers, so the change rate (first derivative) is even more dramatic. It is safe to assume that relatively speaking, few new projects are licensed under a copyleft license. I attribute this to an increasingly enlightened software industry, in which vendors recognize that they can and must work together. To not limit their business models, they predominantly choose a permissive license.

I expect this to keep going. However, I also see another trend on the horizon which I speculate will reverse this trend: User-led consortia.

Vendors develop permissively licensed components for use in their products. Users want applications for use in their business (but not for use in a product). Think Firefox or LibreOffice, not VueJS or apache-commons. Users either develop or sponsor the development of these applications and have little interest in vendor lock-in. Hence, they’ll try to reduce some of it by choosing a copyleft license for the open source applications the development of which they are sponsoring.

Commercial development of open source applications is gathering steam. We measure this by the growth of the number of consortia that are being created to govern the development of these applications. The software industry (vendors and components) is large, the rest of the industry (all businesses and their applications) is much larger.

We call these consortia user-led consortia. If you are a consulting firm and want to organize your customers to join forces and sponsor the development of open source applications for their business, let me know. User-led consortia are an active area of research for us.