> Third, they don't come with the advice / community / expertise of good VCs.
Definitely not business advice/community, but if a startup is doing significant R&D, it can be useful for the scientific and technical advice and community. Not guaranteed by any means, but I've seen it work out well, mainly in the case where a company submits an STTR jointly with a university research group. If the arrangement ends up working, the most valuable thing to the company in that case is sometimes not even the actual NSF grant money, but the connection to the research group it opens up. Having a joint STTR provides an excuse to get a company sort of attached to the research group as an external partner, with a forcing factor ensuring you end up on their schedule with regular project meetings, access to grad students, etc. Can also end up useless, of course, but when it works out well, it can provide value equivalent to something you'd have to pay a lot for if you were hiring technical consultants (especially in areas like machine learning).
Definitely not business advice/community, but if a startup is doing significant R&D, it can be useful for the scientific and technical advice and community. Not guaranteed by any means, but I've seen it work out well, mainly in the case where a company submits an STTR jointly with a university research group. If the arrangement ends up working, the most valuable thing to the company in that case is sometimes not even the actual NSF grant money, but the connection to the research group it opens up. Having a joint STTR provides an excuse to get a company sort of attached to the research group as an external partner, with a forcing factor ensuring you end up on their schedule with regular project meetings, access to grad students, etc. Can also end up useless, of course, but when it works out well, it can provide value equivalent to something you'd have to pay a lot for if you were hiring technical consultants (especially in areas like machine learning).