It doesn't have to generate profit, but it does need to generate more science/better science than the next best alternative way to spend the money. (It's presumably not optimized solely for science, but also engineering and politics, both domestic and international, but it doesn't seem to be the most efficient way to accomplish what it accomplishes.)
The issue that comes up with 'efficiencies' in this context is that much of the money that has been spent at all on ISS-based science would probably not have been spent at all if it had to be done otherwise, because instead of a large up-front cost of the ISS + smaller cost of an experiment to be done inside or just tossed out an airlock, there would be no upfront cost and a large cost for a dedicated science satellite.
This is why Artemis is structured the way it is, the best way to guarantee that the US sticks to a commitment instead of every other president trashing the last one's work is to tie in so many international partners and so much large-scale up-front spending that it becomes too prestigious to cancel.
On top of that, there are lots of engineering benefits, where it's useful that we've learned the lessons from the ISS instead of needing to learn them later at the Moon or Mars or needing to have a sufficiently slow and expensive waterfall design process where they get figured out on the drawing board. Eg. simply being a destination for crewed spacecraft to go to is in itself useful as it creates a standard target to aim for. It allows SpaceX to discover that their toilet can leak urine into a compartment sometimes much faster because they're flying and refurbishing regularly.
It'd similarly be hard to quantify the benefit of things like regular resupply flights needed etc, considering that a big part of what helped SpaceX establish itself as a leader in the industry (and thus also giving rise to all of the other new launch companies and their potential technological improvements, private funding and cost reductions) was the ISS cargo resupply contract they won. Creating that regular demand has a similar effect on funding and costs as the emergence of megaconstellations as a viable project have had.
So, I'd say that ultimately the ISS is proving to have been worth the cost for the US. Without it acting as an easy destination, our space industry would be kinda like the SLS, absurdly expensive, very corrupt, barely usable and essentially non-existent (compared to the rest of the industry).
