It’s a huge design space. One easy to understand example is a passive sonar underwater drone that stays close enough to the surface to gather sunlight and easily communicate with satellites without also being visible on the surface or making sounds from surface wave action. It can then send high frequency radio signals to satellites which rapidly attenuate through the ocean creating a near perfect stealth system vs subs. The option exists to deploy a string of much deeper sensors via a tether.
Another option is close equivalents to the existing systems dropped from aircraft, though with significantly more stealthy deployment.
It's a huge design space but the sea is a harsh mistress. Look at sonobuoys for example; some are passive, just using a hydrophone and relaying any signals via UHF/VHF and others are active where they ping and then relay signals. Generally, they're good for about eight hours. And they have no propulsion system, they just float.
Now you want to add a solar array (without being visible). And easily communicating with satellites will rapidly eat into your power budget, even without worrying about propulsion.
Yes, sensors can be tethered. We've done that since the early 80's, and let our SOSUS systems deteriorate since the fall of the USSR. But these aren't drones, aren't anything like a swarm of drones.
Let's not shift the goalposts away from the original argument.
That’s a function of how aircraft dropped sonobuoys are used. Subs aren’t going to stay still and the ocean is a big place so long lifespans are pointless. However for harbors you want continuous monitoring and so use a different design.
As to talking with satellites, that’s become very efficient at low bandwidth. If your 100W radio is off 98% of the time it doesn’t add up to a lot of energy.
Another option is close equivalents to the existing systems dropped from aircraft, though with significantly more stealthy deployment.
Etc etc.