I'd also like to know about how vigorously major tech companies are enforcing their patents, since they're almost all hoarding quite a few of them. You occasionally hear about some aggressive companies like Sonos, for example, but I'm sure there are many more stories that don't percolate up to the mainstream. Is the only thing protecting startups the goodwill of the behemoths and low visibility? Are there tech startups that end up being destroyed by licensing costs and legal fees related to patent claims?
What are the best strategies for successfully defending against this problem? For example, would it be advisable to search for relevant patents and attempt to license them before working on a project? Should you always simply give up on a project that has patent exposure?
- If you do some sizable project, you will certainly infringe some (or maybe many) patents.
1. Not all infringed patents are defended in trials, at least for one reason, how do the owner knows you are infringing on their patent?
2. If you are a company with deep pockets and facing a small company, you do not risk much by infringing some patents, as you could financially struggle your opponent.
3. If your company does not have deep pockets, or if the patent holder also have deep pockets, you have a problem. IMO the best strategy is:
3.1. Make a patent survey before entering production.
3.2. If the patent holder looks like a reasonable entity (a well known company) try to license the offending patent.
3.3. Do not disclose sensible information to make it harder to detect a patent infringement. For example do not say your modem uses this algorithm, but say you obtained performance the double of your competitors. For a layer this diminishes the possibility that you are using a derivative of someone else's technology as you are much better than them.
3.4. It's always possible to take a patent and change one or two claims in order to tell that what you do, is very different from the original patent.
3.5. In a trial don't count on the judge expertise, or understanding. Don't count on prior art. They will make their decision on points they understand, and they most probably do not understand your technology. The result is often surprising.