At Firefly Aerospace in Texas, engineers have been doing what NASA’s chief, Jared Isaacman, likes to call “the near-impossible.” The Firefly team has been assembling Blue Ghost Mission 2, a robotic spacecraft intended to land on the far side of the moon in 2027.
No U.S. spacecraft has ever gone there. Not very many have landed on the near side, either, for that matter. Blue Ghost 2 has thousands of potential failure points, and there is almost no room for error. Two ships sent by a competitor, Houston-based Intuitive Machines, reached the moon’s near side in 2024 and 2025 but fell over on landing and returned little useful data. In 2024, another mission, Peregrine Mission One, by Pittsburgh-based Astrobotic Technology, missed the moon entirely because a single helium valve wouldn’t seat properly.
In summary: Moon landings are supremely hard. So far, since Project Apollo ended in 1972, U.S. spacecraft have landed intact exactly…once.
And now, NASA, answering a White House executive order to start building a moon base by decade’s end, says it plans something like 70 lunar landings over 10 years—with 45 of them by 2032. Which raises a major question: How will the U.S. pick up the pace so dramatically?
Carefully, that’s how. Carefully, but with all deliberate speed. NASA and its political supporters in Washington frequently cite competition from the Chinese space program, which says it’s planning its own base near the lunar south pole in the 2030s.
Phased Plan for NASA Moon Base
Carlos García-Galán is NASA’s moon base program manager, and his team has laid out a plan with three phases. Phase one, between now and 2029, may involve the near-impossible—putting spacecraft on the moon—but landings need to become regular and low-risk. While the one successful landing so far happens to have been by Firefly (Blue Ghost Mission 1, in March 2025) it’s better to have multiple options. NASA says it expects at least a half dozen companies to try to make landings by the end of 2029. It has called for seven landings in 2027 and another seven in 2028, just for starters.
“We’re not reaching for the home run on the first swing,” says García-Galán. “We have very specific goals, and the first objectives are achieving high-reliability access to the surface.”
If that succeeds, phase two, between 2029 and 2032, would include the first infrastructure—power systems, communications networks, perhaps the first pressurized rover for astronauts, supplied by Japan’s space agency.
Finally, in phase three, astronauts from NASA’s Artemis project would start to spend extended periods at the base. There would probably be fewer but bigger supply missions over time; NASA initially projected it would have to land 4,000 kilograms of equipment in phase one, 60,000 kg in phase two, and 150,000 kg in phase three, with landing ships large enough by then to carry up to 8,000 kg at a time.
A 3D rendering from NASA shows what a near-term lunar base may look like, something the agency hopes to achieve by the early 2030s.NASA
How Would NASA Build a Moon Base Affordably?
NASA’s business approach, for more than a decade, has been to let private companies run actual missions where practical; NASA acts as a “customer,” providing basic funding and sending payloads on each flight. Firefly, Intuitive Machines, and Jeff Bezos’s company Blue Origin already have contracts with NASA, with many more to follow. The more options, NASA says, the better.
“Let’s not invest $1 billion in just one thing,” says García-Galán. “Let’s invest $1 billion in 10 things and see which ones work and use that in phase two.”
There is a survival-of-the-fittest quality to the moon base plan. Companies are given contracts of US $100 million to $200 million to execute moon missions; for the most part, there will be no government bailouts if they run behind or have technical problems. That means there will undoubtedly be winners and losers: In July, NASA and the nonprofit Draper Laboratory agreed to cancel a contract for a landing attempt; the mission had been targeted for 2025 but was running behind by at least five years.
Five years is obviously not good enough when you’re trying to put down roots near the lunar south pole before China does. NASA is pressing companies to get missions going in two years or even less. In June it signed contracts with three companies—Astrobotic, Firefly, and Intuitive Machines—to try to send landing missions in 2028.
“It’s lit a fire under a lot of people to dream a little bit bigger and go a little bit further,” says Kevin Scholtes, a future systems architect at Firefly. “It’s a lot more exciting, what the future looks like, but it’s a lot more hectic.”
On 22 September, Firefly finished a new cleanroom in Texas, four times as large as its existing space, making it possible to run what it called “a dedicated assembly line” and build up to 12 Blue Ghost landers at once. It’s hardly alone.
“I have personally visited over 15 actual or could-be moon base vendors,” García-Galán says, “and every place I go there’s the main building where they’re doing their current work—and then there’s a building under construction next door because instead of being able to make one they’re going to be able to make four.”
Business Successes, and Headaches, for NASA
NASA should be, if you’ll pardon the pun, over the moon about all this activity: If companies from Florida to California are competing for its business, they presumably drive down prices and speed up their work. But it cuts both ways. A corporation may set big, idealistic goals, but it also needs to run a profit. At least one company, Masten Space Systems, was hired for an early landing mission, and then, after four years’ effort, went bankrupt.
A much bigger potential headache is Elon Musk’s SpaceX, the unrivaled rocket giant. Its Falcon 9 makes several launches a week, often carrying other companies’ spacecraft for a fee. But SpaceX, which wants to focus on its giant Starship and its efforts in artificial intelligence, has told people in the space community that it wants to retire the Falcon 9 in a few years.
“We definitely want to move from the older technology to the newer technology,” said Gwynne Shotwell, the president of SpaceX under Musk, at a business summit in Los Angeles on 14 September. “If we don’t obsolete our own products and services, someone’s going to find a way to obsolete them for us.”
That leaves everyone else, including NASA, in a bind. Nobody else launches remotely as often. Blue Origin has been building a super-rocket called New Glenn—but one blew up in May, not only destroying the rocket but its one launch pad at Cape Canaveral. Rebuilding is in progress—with NASA rooting for them because Blue Origin and SpaceX are the two competitors to make astronaut landing ships for Artemis missions.
As people at NASA often say, this is a “high-risk, high-reward” business. But the more businesslike the moon base effort can be, the better. Lunar landing ships historically have demanded the most exacting, mission-specific engineering. If a moon base is to succeed (presuming it’s not canceled outright by a future administration), NASA will need its private-enterprise partners to turn landings into something routine, mass-produced, like a Falcon 9 launch.
Firefly’s Scholtes, for one, says it can happen. He points out that while the first airplanes were handcrafted, within a couple of decades one could buy an airplane as a kit.
But it won’t be simple, he says: “It’s the nature of the space industry that the physics is just barely easy enough to be possible.”