Did the budget balloon and delays rack up? Yup, and if you read about why you will see that it was basically unavoidable. A private business would have canceled it at the first roadblock (depriving us of an INCREDIBLE scientific tool).
And for the actual launch and deployment, the cost of the instrument meant that a very high success rate was very very important. Go ahead and look up the success rate of space X launches and tell me you would put a multi billion dollar tool on their rocket. You need someone who can spend more money, even a lot more money, and guarantee success.
For example, when I worked at Boeing, sinking a die for a forging cost $250,000. Stamping out a forging cost a few dollars. When I was putting together a stack of identical electronic circuit boards, the first one took 2 hours. The last one took 15 minutes.
there are a huge amount of one time costs associated with the design and engineering of a spacecraft or a satellite that could enable fairly cheap constellations compared to singular satellites, and the testing regimen becomes much more cheap when it’s done as part of a campaign.
I’m a proponent for the disaggregation of satellites from sensors using specified interfaces. Industry is definitely moving in this direction with their satellite buses you can buy fairly cheaply.
Will we want to stick a $3 billion sensor onto a $15 million satellite? I can’t answer that question