I don't buy the premise that in this world that there's absolutely no place where a drone can't penetrate, ever. If a drone is large enough to use propellers then it can be detected by its sound and heat signature. If it's small enough to float on air currents then a strong airhandler and dust filter will clean out the drones. Insect size drones can be defeated with lasers!
But assuming that the ubiquitous stealth drones and privacy has disappeared, how would warfare be conducted?
Humans won't call the shots anymore though they will set the rules of engagement. Attacks will happen too fast for a human to decide what to do or even begin to synthesize the situation. Look at high frequency trading on the stock markets now. Humans setup the rules and processes of "engagement" then let the computer run full speed.
Investment in conventional guns and tanks become less worthwhile, though not worthless, because most attacks will be of a digital nature. Why spend 20M on an advanced main battle tank when a $30 drone can hijack the the electronics and render the tank inert? At the same time, investment in the arms race between keeping secrets and discovering secrets would explode. As it is today, some attacks are so expensive, so specialized that they can only be used by nation states. Defense against such attacks takes the resources of nation-states or mega-corps.
Attack vectors: time based attacks, information denial, area denial, info processing denial, drone production facility denial, attacks on info aggregation points, drone subversion.
Time based attacks - This will be the most likely approach. Learning to attack in cyber-space and in real-space so quickly that even someone with perfect battleground intelligence can't stop an attack. For example, knowing that a Rods from the Gods weapon has launched doesn't mean you can do anything about it.
Area denial - Deny an enemy use of a strategic area such as a port or major mountain pass through biological/chemical/nuclear means. This could be especially effective when used against information processing/aggregation points.
Information processing denial - Even if the enemy has total access to your information, if they can't synthesize that information into actionable data, it doesn't matter if they have stealth drones all over the planet. Attack their data centers.
Information mis-routing - The information has been processed but it's only useful if it gets to the people/systems who can actually do something with that info. Route the target info to that goatherder in Afghanistan. It'll be too late before it gets back to the right people.
Drone production denial - With a drone defense system capable of knocking out 90% of an opposing drone force, keeping an enemy from replenishing their supply of in-flight drones would be a huge advantage. Likewise, denying or delaying your enemy access to drone raw materials would decrease drone coverage.
Attacks on drone design - Do what Stuxnet did to Iranian uranium centrifuges by forcing small but critical errors in equipment operation. Inject deficiencies into the 3D printers that make the stealth drones so the drones don't work as well or have compromised stealth abilities.
Defense measures: Drone area denial, Drone hijacking, encryption, greater defense in depth, geographic dispersion, steganography.
Drone Area Denial - While not 100% effective (as stated in the question) even a 90% attenuation will offer sizable information security benefits. Drone countermeasures for everyone! Party fanfare
Drone Hijacking - Make the drones work for you!
Strong Encryption - Remember that the point of encryption is not to keep a piece of information safe forever, but to keep it safe long enough that it won't be useful to an attacker.
Greater defense in depth - Build defensive systems in such a way that compromising one layer doesn't offer the attacker access to the asset. Even better, design the defenses in such a way that if one layer fails, it notifies the next inner layer of the attack.
Geographic dispersion - While uneconomical in many ways for the defender, isolation does mean that if a drone shows up, it's not supposed to be there and is hostile.
Things that won't change much:
- Economic imbalances won't go away. The US would still outspend
the rest of the world to maintain their technical superiority.
- It's really hard to duplicate esprit de corps, even if you know everything there is to know about an organization. This happens in software companies all the time. Rackspace is an excellent example of how corporate culture plays a huge role in how they compete. A competitor may know a lot about how Rackspace culture works but duplicating that culture is very hard. Culture depends on the individuals who make up the group and how they think.