Dumping open source for proprietary rarely pays off: Better to stick a fork in it

LONDON — Over the past decade, companies such as Redis, Elastic, MongoDB, and HashiCorp have abandoned their open-source license roots and switched to proprietary models. The companies cite competition from cloud providers and investor pressure, but the decision is all about making more money. However, there’s a fly in this soup: The strategy doesn’t work. 

At the UK’s State of Open conference, Dawn Foster, director of data science for the CHAOSS Project, unveiled compelling evidence that forks — community-driven alternatives to proprietary codebases — are thriving. At the same time, companies that abandoned open-source principles face stagnant growth and disillusioned users.

Also: Why Mark Zuckerberg wants to redefine open source so badly

The business theory behind going proprietary sounds good. You dump your open-source licenses (for example, Apache 2.0 or AGPL) that got you your start and replace them with restrictive licenses, such as the Server Side Public License (SSPL) or Business Source License (BSL), to more easily monetize your code.

However, the bad news for many companies is the financial payoff has never materialized. RedMonk, the developer analysis firm, published a study by senior analyst Rachel Stevens that found “no clear link between license changes and increased company value.” While revenue grew after the fact, growth rates mirrored pre-change trends. 

At the event in London, James Governor, RedMonk’s co-founder, said: “There is neither a share price rise for public companies nor revenue gains. There’s no clear, ‘Oh, we relicensed and got a hockey stick.’ So, I think that if businesses are making these decisions, the expectation is that relicensing will be the special source that takes it to the next level. The numbers do not indicate that.”

Simultaneously, Foster noted at the event that when companies closed their code, communities fought back with successful forks. For example, OpenSearch, forked from Elasticsearch after Elastic’s 2021 license shift, now boasts contributions from AWS, Logz.io, and Aiven, with 400-plus contributors. 

Valkey, the Redis’s fork, gained traction in just eight days, backed by Linux Foundation heavyweights like AWS, Google, and Alibaba. While the HashiCorp’s Terraform fork, OpenTofu, became a Linux Foundation project, attracting 120-plus contributors in months.

Also: We have an official open-source AI definition now, but the fight is far from over

Foster’s CHAOSS research also revealed that forks under neutral foundations have three times more organizational diversity than their proprietary counterparts. OpenSearch, for example, saw contributions from 45 organizations in its first year — a stark contrast to Elasticsearch’s single-vendor dominance. 

In other words, open-source forks are far more popular than their proprietary counterparts. Foster said users flock to forks to avoid vendor lock-in. Peter Zaitsev, the co-founder of open-source database company Percona, said 75% of Redis users considered Valkey after Redis’s 2024 license change.

In addition, open-source forks iterate faster. Perhaps the biggest open-source fork success story is Grafana, a Kibana fork, which evolved into a $6 billion company by adding metrics support, which Kibana’s parent company, Elastic, had rejected. While Elastic would become a successful company in its own right, this early miscue cost the company billions. 

Foster also noted that open-source forks governed by organizations like the Linux Foundation tend to do well. For instance, OpenTofu‘s governance model ensures no single entity controls the program’s roadmap, making it very popular with developers. 

Also: Red Hat’s take on open-source AI: Pragmatism over utopian dreams

At the same time, forks don’t need developers on board from day one to be successful. Foster said: “If you had asked me a few years ago, I certainly believed this was true when the OpenSearch fork happened; I did not believe that a fork could be successful without forking the contributor community.” She added: “I’m happy to have been wrong.”

Stephen Walli, from Microsoft’s Azure office of the CTO, emphasized the importance of community in these forks: “Projects are not products, despite the fact that a lot of us fall into that kind of mindset.” Too many businesses that start as open-source-driven companies make that mistake. 

So, Foster’s data paints a clear picture: Forks thrive when communities unite against proprietary overreach. Companies that weaponize licenses risk alienating users and stifling innovation, while neutral, foundation-backed forks become the new standard. For investors, the lesson is stark: Betting against open source rarely pays off.