a cruise missile is a UAV that crashes deliberately, and carries an explosive payload.
A UAV that's big enough to carry a grenade or something equivalent, plus the grenade itself, probably won't be much cheaper than a handgun, and it's a single-use weapon. It does have the tactical advantage of being difficult to detect or trace, and of allowing the attacker to be a long distance away when making the attack.
You make a couple of reasonable, but unwarranted, assumptions. You assume that 'grenade' refers to something like an infantry grenade. However in weapons circles, things have a 'CEP' which is the circular error probability that describes the probability that your target will be destroyed based on various distances from the activation of the weapon. In the case of drones, since one could, in theory, manually fly it in very very close to the target, the CEP could be quite small for the device used. Small CEP devices can be made more easily than large CEP devices, and they are lighter.
My thought was to fly a drone up and video my daughters graduation from college, but my wife pointed out to me that it would probably end poorly for the drone (and completely change the tone of memories she should have of her graduation). So I decided against it :-)
Even the cheapest UAV-explosive is more expensive than the explosive without the UAV. A UAV that's "cheaper than a handgun" [0] (as in the grandparent post) is, at present, going to have a payload of perhaps a couple of ounces. How many feet of lethal radius can you get from a couple ounces of payload? How "very very close" can you actually manually fly a UAV with a small bomb attached?
Being able to kill someone with a disposable weapon, and not be spotted by onlookers or security tapes, is still potentially a security nightmare. I'm just not convinced that "cheaper than a handgun" is the present or near-future reality for such a weapon.
[0] cheap handguns are under $250, and rifles can be under $200. Prices taken from http://www.impactguns.com/
The simplest explosive would probably be a single .380 or 9mm pistol round with the primer replaced by an electric igniter. In the field when a gun back fires it often seriously injures the person holding it and people near the weapon. The NATO 9mm pistol round [1] weighs less than 20g and has 600j of energy. Probably wouldn't use the lead bullet. I can imagine (as I am sure others can) an enclosure for this which would not increase the weight significantly but would convert it into an omnidirectional device.
I don't doubt for a moment that you could seriously injure an unarmored individual with something like that within 2 or 3 feet.
The big issue is that 'terrorism' is not about inflicting harm, but fear of harm.
A few quad rotor UAVs outfitted as you describe, possibly even with frangible rounds or just dropping tiny flechettes such as those used from the air in WWI, would easily cause panic.
For that matter, "suicide" UAVs that are either capable of delivering a lightweight shrapnel load without concern for the vehicle's survivability, or just using the kinetic energy of the moving vehicle and the rotating blades to crash into crowds, especially in some sort of flock, would discourage a lot of people from gathering in public places, going to clubs, sporting events, malls, tourist attractions, etc.
Sorry, I'm not from the US and with a cursory google for gun shops found that a Glock pistol of some description (a brand I recognise) is ~$500.
If you have access to basic tools (as in, hacksaw and glue), you can build your own fixed-wing airframe using $100 of foam, carbon rods, and off-the-shelf motors, ESCs, batteries & props & servos[2]. See RCGroups for hundreds of downloadable designs. Your radio control gear will cost you about $40 for the cheap stuff[3]. An ArduPilot 2.5 autopilot is $180 off the shelf [4].
The thing is, if you are building your own scaling up (within reasonable bounds) doesn't really change the price. Something Skywalker sized (~2m wingspan) is easily build-able by almost anyone and will easily take 1kg of payload. $120 dollars[1] gets you a video link & camera with a 20-30km range you can use to fly the thing straight into whatever you like.
So the dumb GPS targeted version is 100 + 40 + 180 = $320
The video targetted version is 320 + 120 = $440
Obviously, that doesn't include the 1kg of 'payload'.
The use of "cruise-missle"-like UAV would likely be different than a handgun. Militaries don't usually use cruise missiles to go after individuals; the asset cost to target value is not proportional in most cases.
Instead, bullets, dumb missiles (target, fire and forget), or (laser) guided missiles from re-usable launch platforms are used instead for those tasks. Cruise missiles are used against larger fixed assets.
I would expect that if "civilian" UAV design evolved toward urban insurgency, the payloads would be used against police stations, radio towers and vehicles rather than individual officers. Getting enough aerial carrying capacity to do disabling damage might not be worth the cost right now, but that equation could change with either better UAVs, or more organized use of urban insurgent tactics.
http://news.discovery.com/tech/gear-and-gadgets/u-s-navy-las...
I think we can expect deployment of smaller systems to police forces as well as civilian drones becoming more and more widespread