"I don't know how to build a $400 million rocket. ... I don't understand how ULA are as expensive as they are."
-- SpaceX COO Gwynne Shotwell

It's been two years since Shotwell fired that quip across the bow of United Launch Alliance (ULA), the rocket-launching joint venture between old-guard aerospace giants Boeing (NYSE:BA) and Lockheed Martin (NYSE:LMT). It's also been two years since SpaceX urged Congress to consider flying its satellites on board $90 million SpaceX Falcon 9s, instead of on $400 million Boeing-designed Delta IVs from ULA.

But as it turns out, the joke is on Shotwell. ULA does not charge the U.S. government $400 million to put a satellite in orbit -- nor the $350 million that ULA advertises.

Instead, ULA charges $422 million.

Spectators watch a Delta IV launch.

Crowds of taxpayers gather to watch 422 million of their dollars blast off into space. Image source: NASA.

U.S. Air Force spills the beans ...

That's the upshot of a new report from Ars Technica. Citing new U.S. Air Force fiscal year 2018 budget estimates predicting that the Air Force will buy three rocket launches from ULA at an average cost of a bit more than $422 million per launch in 2020, Ars calls the new price "complex ... to unpack," and also "strikingly high."

This estimate came to light as a result of a new requirement contained in the 2016 National Defense Authorization Act (the U.S. military budget). Before NDAA 2016, Air Force rocket launches were paid in two parts. One payment was for the rocket launch itself. The other was a $1 billion annual retainer paid to ULA to hold itself ready to launch rockets for the military whenever asked to do so.

Calculating the true cost of any given launch, therefore, required one to add (a) the launch contract itself plus (b) the pro rata portion of the retainer contract, known as ELC, for the evolved expendable launch vehicle launch capability contract. Beginning in fiscal 2020, though, costs are going to become a whole lot easier to calculate. The ELC will go away, and each launch will have its own separate launch price.

... and Ars Technica cherry-picks one bean in particular

Given that SpaceX anticipates doing Air Force launches for as little as $150 million for a Falcon Heavy rocket, or even $80 million for a Falcon 9, those ULA launch costs are going to look pretty steep relative to SpaceX's pricing come 2020 -- but maybe not quite as steep as the U.S. Air Force budget implies.

Responding to the Ars article on Twitter, ULA CEO Tory Bruno complained that the $422 million price cited in the article was a "cherry picked odd number" -- inasmuch as it applies only to launches in the year 2020, and ignores lower estimated prices in other years. In the infographic he tweeted and in his replies to Twitter followers, Bruno further highlighted ULA's own calculations that, across its entire fleet of Boeing-designed Delta IV rockets and Lockheed Martin-designed Atlas Vs, the average cost of a rocket launch at ULA was only "$225 million ... per mission" -- and that was including the pro rata costs of ELC.

That may or may not have been the case historically. But if you look at all the estimates contained in the Air Force's report, it appears Ars actually isn't too far off the mark regarding launches scheduled for 2020 and beyond. The $422 million price cited for 2020 is expected to rise to $423.6 million in 2021, and then fall to $392.1 million in 2022.

The average cost of the eight launches covered over the three years, though, is still $411.1 million -- which really isn't too far off from the "cherry-picked" $422 million number, nor from the $400 million price tag that SpaceX's Shotwell mocked back in 2015.

Mysteries abound

A bigger question is, what happened to ULA's efforts to drive down the cost of its space launches?

Recall that in past statements, Bruno has said that ULA's ultimate goal is to cut its launch costs to as low as $99 million per launch once it gets its new Vulcan rocket into production. Vulcan is expected to begin flying in 2019, which one would think allows plenty of time for the Air Force to factor it into its 2020 estimates.

What's more, even before Vulcan gets up and running, we've observed ULA offering steep price declines on Air Force launches -- Atlas V rockets falling in price from as high as $190 million to just $146.5 million, and $350 million Delta IV rockets discounted to $270 million. While neither of those prices quite reach ULA's ultimate goal of a $99 million price tag, they do show prices coming down as much as 23% already.

Yet whether you key in on the $411 million average cost of launches the Air Force is budgeting for, the $422 million "cherry-picked" figure for 2020, or even the lowest $392 million cost of launches scheduled for 2022, none of these figures show anything close to a $99 million launch price. None of them even shows a 23% discount from current prices. In fact, none of them suggests prices very much different from the $400 million Shotwell accused ULA of charging two years ago.

The big mystery here is why the Air Force doesn't seem to think ULA will cut its prices much at all over the next five years. Investors in Boeing and Lockheed Martin had better hope the Air Force is wrong about that, though. Because if those prices don't come down, I fear SpaceX is going to steal a lot of business from ULA -- and unless it cuts its prices to compete, ULA may not remain in business much longer.

Rich Smith has no position in any stocks mentioned. The Motley Fool has no position in any of the stocks mentioned. The Motley Fool has a disclosure policy.