You're probably right from an Aurora-as-video-game perspective: players hate setbacks. Sid Meier's autobiography discusses this: the Civilization team considered implementing a "civil war" game mechanic, but eventually shelved it because most players hated losing their cities. Even Paradox Interactive's grand strategy games, which do feature civil wars and rebellions, tend to make them relatively minor affairs. There's probably some sort of loss-aversion psychology going on here.

On the other hand, I think that random losses are useful from the Aurora-as-storytelling perspective. The only way for the player to experience "The Glorious Liberation of Hiigara" is as the sequel to "The Hiigara Disaster".

I think that a middle ground could be for the Precursors to only launch their ships after winning the ground war (perhaps there could be an escalating series of ground awakening). That way the player can recapture their (mostly) intact colony later.

Or just use Spacemaster to kill the Precursor fleet. Aurora is a single-player game; there's no such thing as cheating.

Difference here is that resources in aurora are limited (well, besides never ending expansion, and perhaps some exploits, if they can be called that).
In Civilization and Paradox games, resources are more about territory controlled, rather than something finite. The only finite thing that player loses there is time.

You can at least prepare for the coming setbacks in PDX games though - like disasters having specific requirements to occur, and to end.
For unhappiness in Civ games, you expand to places with special resources to compensate. Or, use technology and buildings. There are ways to address losses.
But here though... I don't see any clear way to prepare for this, besides amassing a fleet "in case they wake up". And even then they get the drop on you, because they don't suffer a jump shock or it's equivalent.
Or, simply deciding to not bother going to this system to begin with. Which is... arguably questionable game design. The player should be encouraged to explore and expand. Why this should be an exception?

Basically, I would be fine with this challenge, if there were ways for player to address it. But, I don't see it, hence my concern.

EDIT:
And yeah, making fleet appear only after the ground war is won does look like a nice compromise.

EDIT:
At the very least, my suggestions about odd happenings as your excavation teams dig stuff up gives an indicator to player that something is afoot. The player can at least decide to reduce potential loses and evacuate things.
As for the excavator work - perhaps it would be better to have several different outcomes? Like, in terms of ships:
Outcome 1 - excavator team finds the precursor ship intact. They reprogram it, and get it online (if you have population on the world, you also get it crewed).
Outcome 2 - excavator finds a non-functional precursor ship. You get resources/components.
Outcome 3 - excavator team accidentally reactivates precursor ship. Luckily, it suffers a serious malfunction. Spawn a wreck in orbit.
Outcome 3.5 - excavator team accidentally reactivates precursor ship. It suffers a minor malfunction. Spawn a damaged precursor ship in orbit.
Outcome 4 - excavator team accidentally reactivates precursor ship. It is now in orbit and does what is described in this campaign.

Precursor land forces:
Outcome 1 - excavator team finds remains of a precursor army. You get some resources and maybe research points in some places.
Outcome 2 - excavator team accidentally reactivates a small group of precursor forces. Spawn a small group of hostile forces on planet.
Outcome 3 - excavator team accidentally reactivates a precursor group, which in turn reactivates their army. Spawn a large precursor ground force.
I am not sure if player should be allowed to reprogram robot armies..?

Seeing that precursors will just spawn out of nowhere - in orbit of the planet, with giant ships no less - after digging it up a bit, I think I'll turn the 'enhanced precursors' setting off.
I'll probably turn off for all of the special NPRs, as that looks extremely annoying. Unless I decide I want some masochism for some reason...

You put expensive equipment on the planet, only for it to be completely obliterated by an enemy appearing in orbit - with no downside for them, like jump shock.

If anything, this basically prompts me to ask for a similar tech for players and regular NPRs - fleet teleportation installation. This installation would allow the player/NPR to basically teleport things between teleport installations, limited by installation size. I think size 1 installation would allow 100-ton ship teleports.
But of course, the precursors are darn robots which can teleport, and the pathetic organics can't! I mean, that would be my justification for why regular NPRs/Players wouldn't be able to use it.

You're probably right from an Aurora-as-video-game perspective: players hate setbacks. Sid Meier's autobiography discusses this: the Civilization team considered implementing a "civil war" game mechanic, but eventually shelved it because most players hated losing their cities. Even Paradox Interactive's grand strategy games, which do feature civil wars and rebellions, tend to make them relatively minor affairs. There's probably some sort of loss-aversion psychology going on here.

On the other hand, I think that random losses are useful from the Aurora-as-storytelling perspective. The only way for the player to experience "The Glorious Liberation of Hiigara" is as the sequel to "The Hiigara Disaster".

I think that a middle ground could be for the Precursors to only launch their ships after winning the ground war (perhaps there could be an escalating series of ground awakening). That way the player can recapture their (mostly) intact colony later.

Or just use Spacemaster to kill the Precursor fleet. Aurora is a single-player game; there's no such thing as cheating.

Does the AI share the information across all instances of that spoiler?  IE, will any future Necrons in other systems know the details about the player's ships performance gleaned from the battles in Orpheus?

The AI shares information on a racial basis. So all Necrons will know the same information, but different swarm races won't share with each other.

There is mention of keeping the shields down to withhold that intel from the spoilers. Is that just RP, or are there behavior changes from them if they get that info? i.e. changing missile salvo densities and narrower targeting concentration.

The AI tracks information about your ships, including observed point defence performance and how much damage they took to kill in the past. They will include shield strength in that calculation when determining how many missiles to allocate to a given target.

However, in this case I was keeping shields down to deny them sensor information on the location of my ships. Active shields have a strong signature that can be tracked by EM sensors.

Great to see a new Steve campaign, and the new AI seems to make the Necrons a proper threat. I assume that they are similarly more difficult in their non-"advanced" traditional form, maybe even more so since they will be able to maintain their missile stockpiles. Very happy to see the AI learning better ways to deal with fighters and bombers, although I do wonder if they'll be able to handle more traditional bomber strikes from much longer range (40+ m km versus ~10-20 m km).

Hopefully as this campaign goes on we will see combat with a few different NPRs and most of the other spoilers as well. 

Steve, is there a setting to change the size of the spoiler ship classes? Or is it random and you've just had good luck here?

I'm curious how long you play a given campaign for, Steve. I find that long-lived Aurora 4X let's plays are few and far between, for better or worse, and that mostly because the nature of the game means that complexity and micromanagement ramps up quickly as the number of systems and colonies grows. I saw the screenshots of the galactic maps you posted in the Changes thread showing 100+ star systems in a single game. I don't know how you do it.

This is shaping up to be a very interesting campaign! Some very exciting new additions are noticeable, enemy trigger discipline against small targets and single fire control use in particular. And missiles with active sensors and retargeting? Don't believe I've ever seen that in the current version!

However, the Necrons had destroyed their ordnance stockpile once they knew defeat was inevitable.

Maybe they could stop doing that, now that we're finally getting a use for it beyond scrapping it all? Leaving their entire missile stockpile for us certainly wouldn't do, since it's seemingly limitless, but with an entire planet's worth of ordnance even the Necrons could miss a few dozen missiles here and there...

Five seconds later, a further four Furies were destroyed, although they were able to return fire from 37,500 km and inflict over fifty strength-1 hits on the Khopesh, including fourteen that penetrated its armour. The third volley from the Khopesh killed four more Furies, before it was buried under an avalanche of lascannon fire, suffering forty-five internal hits. The Necron fleet chose that moment, with the fighters at point-blank range, to launch sixty-one anti-ship torpedoes.

Did the enemy ships conserve their ammunition until they realized the danger presented by a swarm of fighters at point-blank range? If so, this is seriously impressive, but I have to confess I don't really understand this part, as the fighters' weapon capabilities should've been known to the Necrons from all the point defense fire they've observed.

Also, I can't help but ask, given your efforts to ambush the Thokt and the Thanatos group in the end, does the AI still have no concept of lost contacts/object permanence, or did it deem the inner system safe because a few months passed between the two fights? Really hoping for the latter, but very excited for the new update either way!

The AI changes seem to have significantly raised the difficulty, the Necrons out-teched you but even so that was a pretty gripping fight. The new version can't come soon enough for me.

One note though, the Inquisition was formed after the Horus Heresy, during the Great Crusade and Heresy era intelligence was handled by ad hoc groups from virtually all human factions working at cross purposes. Which I recognize might be hard to represent, just the 40k pedant in me.