In the early days of computing, there were no software products. Some computer systems vendors provided mechanisms for their customers to freely distribute applications with their source code, typically through their user groups (e.g, IBM SHARE Library) [1]. Applications were custom-built, developed across many different computer systems using a wide variety of different languages.
The 1964 arrival and subsequent broad adoption of IBM’s System/360 and OS/360 made it feasible for developers to build software products that could be licensed to a large number of buyers. Soon thereafter, Martin Goetz was able to patent a software algorithm and to release Autoflow, the first commercial software product. Those steps led to today’s multi-billion dollar commercial software industry, with vendor-licensed products, accompanied by End User Software Agreements and patents.
While vendors focused on enterprises and governments, the arrival of personal computers in the mid-1970s, created a consumer market for software products. In addition to vendor supported products such as Microsoft BASIC, other programs, collectively known as “shareware”, were developed by individuals, with their users asked to make a small donation (under $50 US) to the developer. Some of the most popular applications of this category were PC-Talk, PC-File, PC-Write, and WinRAR, all of which attracted many thousands of users. Shareware remained very popular into the 1990s.
There were many different commercial approaches to shareware. Some applications were released for trial use, and would work for a while, often a month, before asking for payment. Those who paid would receive a code (key) that would allow the program to keep working or perhaps activate some hidden features. Some shareware worked on an honor system, leaving the program operating normally, hoping that users would make a donation to the developer, but not requiring it. There were numerous variations of shareware commercialization, and several services were created to help developers collect payments. Shareware programs were released in binary format, without the source code to modify them or to see how they operated.
In the late 1970s, the University of California, Berkeley, released binary and source code for BSD Unix on the DEC VAX 11/780, using a new type of source code license. Academic institutions could obtain the software for a distribution fee, while companies had to pay more than $40K for a commercial license from AT&T. BSD Unix was widely adopted in universities, and led to the development and distribution of many applications. My group released several applications in 1980 with BSD licenses, making them early examples of open source software.
In the mid-1980s, though, Richard Stallman created the Free Software Foundation and the GPL license, leading to distribution of his team’s GNU software in both binary and source code formats. When Linux was distributed in the early 1990s, it followed this non-commercial approach, using the GPL license. The concept of “free software” and the success of Linux following its 1992 release led many developers to follow the same approach.
All of the above developments influenced the growth of open source software. The MySQL project, begun in the mid-1990s, used a dual license model, which had the advantage of allowing developers to use the open source license until they were ready to release their product commercially, at which point they had to switch to the commercial license. The code was unchanged, without the additional “open core” components that are common in today’s Commercial OSS (COSS) product. When MySQL was acquired for $1B in 2008, it showed the value of COSS, which made the category attractive to private investors and VCs, an important step in the evolution of COSS.
By the early 2000s, following the creation of the OSI and the release of the Open Source Definition, many projects had identified ways to generate revenue from their open source projects. I wrote a paper in 2007, “Building a Business on Open Source Software” [2], that identified different approaches. These included providing support and training services, hosting the software for remote access, commercial enhancement, and integration with other products. These approaches to commercialization remain the primary means by which COSS companies generate revenue, as described in the Monetization section of The State of Commercial Open Source 2026.
It’s valuable to see COSS as just one approach to open source software and to recognize that there remains a very large community of open source users and contributors who work on non-commercial applications, including both community-based development and foundation-sponsored projects. Development processes and team participation in these categories are different from those of COSS and proprietary software products, as described in “Community and Commercial Strategies in Open Source Software” [3], published in 2013. Many of the projects In the former category give little attention to monetization for their work, and volunteer their time and resources. There are more than a billion (!) projects hosted by GitHub, many of which are non-commercial. While it can be difficult for volunteers to contribute to COSS projects, the non-commercial projects are often seeking such help.
As the Linux Foundation report describes, the role of AI in COSS is still uncertain, with major questions revolving around the use of proprietary AI coding tools in the development and maintenance of COSS. One can also foresee non-commercial open source software produced by AI tools in addition to the existing contributions of AI coding tools to existing open source projects. These topics are likely to take on growing importance as AI technologies evolve.
In conclusion, the idea of creating a business based on open source software has evolved for decades, going back to the 1960s and the first instances of software being distributed independently of the hardware on which it ran. Today, COSS represents an important segment of software business approaches, though there remain significant questions about the openness of many COSS projects and the licenses under which some of them are distributed.
Tony Wasserman
Software Methods and Tools
[1] SHARE (computing). https://en.wikipedia.org/wiki/SHARE_(computing), accessed 6 November 2026.
[2] Wasserman, A.I., “Building a Business on Open Source Software” in Petti, C. (ed.), Cases in Technological Entrepreneurship: Converting Ideas into Value. Cheltenham, UK and Northampton, MA, USA: Edward Elgar, 2009, pp. 107-121.
[3] Wasserman, A.I., “Community and Commercial Strategies in Open Source Software”, IT- Information Technology, vol. 55, no. 5 (2013), pp. 181-188.