NASA is buying cargo delivery, and SpaceX is using it as a lunar factory permit

NASA has selected eight companies — including SpaceX, Blue Origin and Lockheed Martin Space — to compete for lunar delivery work worth up to $10 billion over the next decade, under its Commercial Lunar Cargo Services programme. The contractors would build and operate landers carrying scientific instruments, technology and infrastructure to the surface.

The contract is the concrete part. What SpaceX says it would do with that access is the part worth separating out.

$10B

ceiling for the lunar cargo contracts, competed over the next decade

companies selected to compete, SpaceX and Blue Origin among them

$1B

already committed to Blue Origin, Impulse Space and Arrow Science and Technology for separate lunar missions

What NASA is actually buying

The agency's permanent-base plan runs in three phases, and each phase is a procurement.

Exploration — to 2029
Robotic rovers survey the lunar south pole to identify viable sites for future infrastructure.
Construction — 2029 to 2032
Power systems, cargo transport and a communications network, installed to support long-duration operations.
Habitation — after 2032
Living space completed so crews can run larger research programmes and travel further into the solar system.

The cargo contract sits under phase two. Nothing in it pays for a factory.

Delivering mass to the Moon is the bottleneck that pushed NASA toward commercial suppliers. The same agency has also selected SpaceX's Starship to launch satellites in 2028 to test communications technology, and awarded contracts — including to Blue Origin-owned Honeybee Robotics — for power generation and wireless power transfer.

Lunar cargo contracts, ceiling10$ billion

Already committed to three suppliers1.3$ billion

Two different pools of money, as reported: the competed cargo ceiling and the separate missions already funded.

What SpaceX says it would build

In its S-1 filing ahead of a public listing at a $1.77 trillion valuation, SpaceX describes building factories on the Moon that manufacture AI satellites from lunar raw material, so that only the chips have to be shipped from Earth.

That is a supply-chain argument, and it is the strongest part of the pitch: launch mass is the cost driver, and a satellite built where it will operate avoids launching its own structure and shielding.

Get the equipment there
Every kilogram of plant has to land first — the cost the cargo contract is meant to spread across a decade of flights.
Generate power on site
A factory cannot run on imported fuel; NASA is separately funding power generation and wireless transfer, including at Blue Origin-owned Honeybee Robotics.
Stay in contact with Earth
The 2028 Starship satellite launch is a communications test, not a factory demonstration.
Then build from local material
The economic case rests on replacing imported mass with lunar material, so that only chips travel from Earth.

The filing’s premise in one line: the factory is only worth building if the launch mass falls to chips.

Elon Musk put the ambition in numbers at an investor meeting on 4 August, as reported: a robot-operated AI satellite factory "sounds like science fiction but will happen", and combined with solar power and radiation management the site would let humanity scale AI in space to "1,000 times the Earth economy, even a million times".

The part that is not absurd

Patrick Suermann, a professor of construction science at Texas A&M, told reporters the idea is not unreasonable — and the physics behind that judgement is simple. The Moon's gravity is roughly one-sixth of Earth's, so launching a satellite from the lunar surface costs a fraction of the fuel of launching it from Earth.

SpaceX's filing adds three mechanisms: lunar material replacing imported mass, a mass driver that would use electromagnetic force to throw payloads into space without a rocket, and exports of lunar minerals — the company estimates the surface holds more than a million tonnes of "rare materials" usable in quantum computing and nuclear power.

Whether any of those work at industrial scale is unknown. What is known is why the company wants them: an AI compute constellation in orbit, which would let it add capacity without building more energy-hungry data centres on Earth.

Claims, plans and contracts

Three different things are being reported as one story.

The contract is real and bounded: a $10 billion ceiling, eight competitors, decade-long delivery task orders. Nobody has been guaranteed$ 10 billion.
The factory is a statement of intent in a prospectus — a document written to describe a business to investors, not an engineering schedule.
The multiplier is a claim by the founder, not a projection anyone has published a method for.

The filing's own framing is more modest than the quote: the Moon as "the first large-scale industrial space economy", and over time a strategic industrial and transit hub. That is a thesis about positioning, and it survives without the million-fold figure being true.

Why this matters beyond space

The pattern is familiar from terrestrial infrastructure: a government buys a service it needs, and the supplier uses the contract to build a capability with a much larger private market behind it. NASA needs cargo delivered; SpaceX needs a reason for a lunar supply chain to exist.

Both can be satisfied by the same contract, which is why the award matters more than the announcement.

Source: VnExpress, 13 November 2026, citing Quartz, Fortune and Scientific American. NASA's three-phase lunar base plan, the eight-company cargo selection, the $1.3 billion already committed, the 2028 Starship communications test and the S-1 details are as reported there. The$ 1.77 trillion valuation and the 4 August investor remarks are as described in that report; Musk's multiplier is his own estimate. No engineering schedule for a lunar factory has been published.

Contact: [email protected]