The Prime Exits: How Boeing, Lockheed, and Northrop Walked Away From NASA's Commercial Space Bet

Blue Origin's New Glenn rocket aced its booster landing on April 19. The upper stage left a communications satellite in an orbit too low to operate — the first major failure of a rocket NASA is counting on to help return astronauts to the moon. The booster works. The rest of the mission didn't. That gap is the iterative hardware model in miniature: you fly, you find out what broke, you fix it, you fly again. NASA is now relying on this approach for the most demanding mission in decades, and the evidence suggests it functions for one company.

SpaceX has run 46 missions to the International Space Station since 2020, demonstrating the approach at a scale no other commercial provider has matched. Boeing couldn't execute it. Blue Origin just failed on a component that should have been elementary. And the traditional primes — Lockheed Martin, Boeing, and Northrop Grumman — have told NASA in clear terms they will no longer bid fixed-price contracts. Not because the model is broken. Because they cannot win in it. (Ars Technica)

SpaceX bid $680 million for the US Deorbit Vehicle. Northrop Grumman, the only other serious bidder, came in at roughly twice that price. SpaceX scored 822 out of a possible 1,000 on NASA's mission suitability evaluation. Northrop scored 589. NASA's own source selection statement described the competition as "a total stomp." (Ars Technica) NASA's deputy administrator and associate administrator both told colleagues internally they did not want SpaceX to win without competition. The competition happened anyway.

The primes said the same thing about cost-plus before Starliner. Boeing had all of it — guaranteed costs, shared development burden, agency engineers embedded in the program. It spent $4.2 billion of NASA's money and delivered a vehicle NASA classified as a Type A mishap, the most serious failure category. Boeing has lost more than $2 billion on the program and is exploring selling the whole thing. (Manhattan Institute)

NASA's inspector general estimates the agency will have spent $93 billion on the Artemis program between 2012 and 2025. Each SLS launch will cost $4 billion, is expendable, and is based on shuttle-era technology. SpaceX is developing Starship as the Human Landing System provider under a fixed-price contract for roughly a tenth of that per-mission cost. Whether Starship works at lunar scale is the open question. SpaceX's revenue was $13.1 billion in 2024, up from $8.7 billion the prior year — but revenue is not a measure of whether the hardware will function when a crew is on board. (Manhattan Institute)

The question is not whether the model works in theory. SpaceX has demonstrated it works in practice. The question is whether it survives when there is no competition. Every other bidder has dropped out, fallen behind schedule, or lacks operational hardware. NASA has one qualified provider for its most critical programs. That provider is also its primary launch provider and the dominant commercial operator in low Earth orbit. The model SpaceX proved possible — cheap, frequent, iterative, American — is running into the constraint that it was proved by one company, for one company, on terms nobody else has met.

"We really underestimated how much people did not care what the end-state goal was," former NASA deputy administrator Lori Garver told the Manhattan Institute. "They really wanted to make sure the money would flow to those contractors." (Manhattan Institute)

The money is still flowing. The contractors are not. And the government that bet its human spaceflight program on a commercial model only one company can deliver is discovering that operational redundancy and competition are different things.