SpaceX Pulls Back From Government Launches as AI Takes Over Revenue, Rattling NASA and Space Force

SpaceX (SPCX) is stepping back from new U.S. government launch business as it shifts toward artificial intelligence, according to reports that have alarmed officials at NASA and the Space Force, which rely on its low-cost, dependable launches. SpaceX has told NASA it plans to stop flying its Crew Dragon astronaut capsule by or before 2030, has stopped offering commercial Falcon 9 launches beyond 2028, and declined to offer a rocket for three upcoming Space Force missions. Existing government contracts remain in place.

What is changing

Crew Dragon and NASA

SpaceX President and COO Gwynne Shotwell has said NASA should give Boeing some business and fly the capsule the government paid for, pointing crew transport toward Boeing's Starliner. The timing is notable: on Sept. 18, NASA awarded SpaceX $946 million for three more crew missions (Crew-15, Crew-16 and Crew-17), lifting the total to 17 missions and $5.92 billion, with a period of performance through 2030. NASA has said it wants two commercial crew providers. Boeing has yet to fly Starliner on an operational space station mission, and NASA has reportedly committed additional money and ordered two more Starliner crewed missions in response.

Falcon 9 and the Space Force

SpaceX has stopped offering commercial Falcon 9 launches for missions after 2028 and may eventually stop flying NASA missions. NASA's planned ISS deorbit vehicle could need a different rocket if it launches after 2030. The Space Force sought rides for three future National Security Space Launch missions, two of them for the Space Development Agency's low-Earth-orbit satellite network, and SpaceX declined all three; no other provider offered to take them. A Space Force official said the service is actively talking to SpaceX about when Falcon missions will end and acknowledged "growing pains" as demand rises.

Alternatives are limited. United Launch Alliance's Vulcan was grounded earlier this year over solid-rocket-booster problems, a Blue Origin New Glenn exploded on its only Florida pad in May, and many military satellites are not built to fit SpaceX's Starship, which is designed for flat-stack satellites such as Starlink.

Why: the business is moving to AI

SpaceX's third-quarter revenue was about $14.58 billion, of which AI contributed $7.56 billion, or 51.9%, Starlink $5.25 billion (36.0%) and rockets $1.77 billion (12.1%). In the second quarter, AI took about $15.8 billion of $18.4 billion in capital spending, versus $1.2 billion for space. Of 113 launches in the first nine months of 2026, 84 were internal, such as Starlink, so launches for outside customers are a small slice of revenue. Management increasingly expects AI, not traditional launch contracts, to drive growth, while military data services such as Starshield remain attractive.

What it means

The government built its launch plans around one provider that flew most U.S. missions at a lower cost than rivals. If SpaceX steps back faster than alternatives mature, NASA and the Space Force face higher prices and fewer launch options in the late 2020s. That could favor companies with competing capacity, including Boeing (BA) through Starliner, Lockheed Martin (LMT) and Boeing through United Launch Alliance, and newer entrants such as Rocket Lab (RKLB), though none has yet shown the cadence SpaceX has. For SpaceX investors, the shift is consistent with the company's own mix: AI and Starlink are growing, while government launch is a small and shrinking share. These are general observations, not forecasts.

Sources: NASA announcement of the SpaceX crew contract modification (Sept. 18, 2026); SpaceX third-quarter 2026 filing. This is informational, not a recommendation to buy or sell.