I agree that Bird likely can't make its unit economics work with its current model, but if its goal is to be acquired by a company who wants the user base, then $2.5B isn't nearly as crazy.
Remember that Bird has two kinds of users, both of whom are monetizable by other companies: 1) riders, 2) chargers
But any potential buyer can build his or her own and get the same level network effect. Bird doesn't have a network effect. You know this because you don't go out looking for a Bird. You go out looking for a scooter. Or you buy your own scooter. That's why Bird won't be bought, and why it's probably worth $100M max.
It's not network effects that matter. It's economies of scale. Any company can go out and start putting scooters on the street. But if you want positive unit economics, which Bird now has for over 2/3 of its fleet, you need a huge infrastructure.