(computer science perspective here)
Research in grad school differs from industrial research in a few key ways. First, research in academia can be more abstract - you can focus on fundamental research and not feel pressured to produce, say, a working prototype or a product.
Second, the type of research you can conduct is dictated by company needs: if your company is interested in deep learning then you're gonna work on that even if you're really interested in graph theory.
Third, since companies have an incentive to provide products, you may be under more pressure to produce within tighter timeframes.
Finally, often enough companies limit your capacity to publish results due to confidentiality/company interests. Companies won't like it if you publish results on a new method for competitors to use (you may not publish anything for years if you sign an NDA), or publish things that present the company in a negative light (Timnit Gebru's ousting from Google is a notable recent example of such a conflict).
There are of course exceptions. Some industrial research divisions offer more academic freedom, but even then you are expected to contribute to the company's bottom line (which is not unreasonable in my opinion).
That said, companies have a clear advantage when it comes to access to resources. Generally speaking, academia simply cannot compete with the resources and datasets that tech giants such as Google/Amazon/Facebook/Intel/IBM have at their disposal (I don't know enough about other fields, e.g. how the NIH compares to big Pharmaceutical companies).