Yes. Letting them start over on something completely different is indulging sunk-cost fallacy. People were hired on the strength of being able to build X. If they are not who you would have picked to build Y, then getting them building Y anyway means you start out with a broken leg.

Not sure about this, it's very contingent on the specifics of X and Y as well as how specialized the hires were. Obviously it depends on the domain, but I definitely bias towards generalists of the Joel Spolsky variety (ie. "smart and gets things done") versus loading up on specialists early as it can often constrain the thinking in a way that hampers exploration when you're trying to find product market fit.

On the other hand, if the team works well together, maybe they would actually be good at building something different?

It seems like this will vary significantly depending on the team and the new plan. How valuable is it, really, to keep this team together?

As investors, we have no more knowledge of how well the team works together than that they failed to build what the market wanted. That is their only track record as a team.

It's true that from the outside view, we know very little. Insiders may disagree, though.

that's not really sunk cost fallacy? sunk cost fallacy in that case would be not pivoting when it's the best move looking forward, because you already spent so much time building and hiring for x (i.e. the sunk cost, which if not falling into the fallacy you should be ignoring for that decision)

You spent a potful of money, and got nothing. You can give the same people another potful of money, or give it to somebody else who demonstrates actually delivering something of value. Only sunk-cost fallacy says to give the former, exactly, another potful. They should get back in line with everybody else.