Risks and opportunities of open-source are misjudged
When talking with people of different backgrounds, the common pattern is that they fear someone will steal their ideas if they open source their code.
I can partially understand this if, for example, you release the code to analyze data before you finish analyzing your dataset. Smart people could guess what you are trying to do and scoop your research . First, I don't know if this fear is justified, especially not for experimental scientists. assumption
If you are a scientific hardware company, I have heard people saying that by open-sourcing they enable competitors to build on what they developed and invested on.
I can also partially understand this, but only in the cases where software is your core business, and not the hardware itself. Imagine you provide data analysis services as a subscription. However, the reality is completely different if you are really a hardware company.
If you are a researcher, you can generate a larger impact through your software than through your papers. Software can be readily used, cited and speed up the development of science. It also helps addressing the reproducibility issues. If you are in a non-stable career, it helps to show your abilities to future potential employers, much more than a paper.
Also, collaboration can be boosted through software. Contributions are properly acknowledged, and the evolution is documented.
For companies, the discussion is completely different. Open-sourcing can have an economic impact by lowering the requirement for licensing. It also provides a path to lower the costs on support, since users could solve their own problems. Moreover, it is a great channel to capture value. Most users won't ever care about the software being open-source. But some power-users can actually provide improvements, thus improving the product for everyone at virtually 0-cost for the company.
This is without even considering the possibility of creating a community around the software released by the company. There's nothing stronger than a group of committed users. Of course, there are all the other moral considerations regarding how it should be, especially for scientific hardware.