Costs of the robotic Hubble repair mission have been skyrocketing.

The estimated price tag of a robotic rescue mission — between $1 billion and $2 billion — is raising eyebrows and questions about whether Hubble is worth the investment amid tight budgets and periodic reports of technical woes that could cripple the spacecraft before the robot gets there.

I’ve never taken this mission seriously. I don’t think that NASA ever really intended to do it. The initial studies were just a fig leaf to distract attention from the fact that they weren’t willing to send a Shuttle to it, and assuage Hubble fans. The problem that they have now is that just safely deorbiting the thing is going to be impossible to do for a reasonable amount of money. I still think they should do the Shuttle servicing mission, because the marginal cost of that is the absolute cheapest thing they can do, and the risk is overblown (though even if it’s as dangerous as some think, it’s still one of the few things that Shuttle could do that would actually be useful).

By the way, they (like almost everyone) gets this part wrong:

If the cost hits $2 billion, that’s three to four times what it would cost to send astronauts to do the job as they have four times before and as NASA planned before the Columbia disaster.

That’s not what it would cost to send the Shuttle. The marginal cost of a Shuttle flight is somewhere between a hundred and hundred fifty million dollars. They’re basing this assessment on the average cost, which is more than half a billion, but that’s not the number one would properly use to make that decision.