Sharing results that are not commercially viable would speed up research
I once worked on an industrial research project that aimed to develop a carbon fibre reinforced polymer composite with exceptional mechanical properties. After spending around two years working on different formulas and generating a ton of results, it started to become obvious that the envisioned targets were still far out of sight. A decision was taken to cut the losses and scrap the project.
If the target mechanical properties could have been realised, the discovery would have been a major commercial success. Yet, now, after years of time and money spent, it had no financial viability as a patent and had no place to fit within commercial research infrastructure. So, the knowledge generated all went into (eternal) storage, never to see daylight again.