Astro Teller has a test he runs on rooms full of executives. Choice A: add a million dollars of value to your business this year, guaranteed. Choice B: add a billion dollars of value, but with only one chance in a hundred. Teller, the co-founder of X, Alphabet's innovation lab, whose title there is Captain of Moonshots, said almost nobody anywhere picks A. Nearly everyone picks B, and Teller tells them they have "passed the math test": a 1-in-100 chance at $1 billion is worth $10 million on average, ten times the guaranteed option.
Then comes the second question. He said he has asked "CXOs the world over" to keep their hands up if, on their best days and in their wildest dreams, their manager, CEO or board of directors would actually support them choosing B. Every hand goes down.
"You don't need a lecture on innovation, you need a new manager," Teller tells them.
Teller told the story during a live interview at Moonshots Live 2026, recorded on September 25, 2026 and released as the 300th episode of Moonshots with Peter Diamandis on October 5. Over the course of the conversation the hosts kept pressing on one question: if big organizations say they want moonshots, meaning ambitious long-shot projects that aim at huge improvements rather than small gains, why do so few of them actually take any?
Moving the long shots to the edge
Peter Diamandis, the founder of XPRIZE and Singularity University, set up the hand-raising story by summarizing what he said he had learned from Teller, whom he called his "moonshot mentor." A company's core organization is responsible for delivering 10% profit year after year. The moonshot group belongs at the edge of the organization, where the "crazy thinkers" work. Its job is 10x improvements, not 10%, and it should report directly to the CEO. Diamandis paraphrased Teller's own rule for that group: "if you work on anything at 10%, you're fired."
Teller said the formulation was "exactly right," and explained the reason with his experiment. The choice-B projects have to be sequestered, he said, "to a place where they're actually wanted, and the mess that comes with them is tolerated."
Diamandis offered two historical parallels: Lockheed's Skunk Works, its separate advanced-aircraft unit, and Steve Jobs taking the Macintosh team off campus to build the computer. The Mac example is well documented. Macintosh developer Andy Hertzfeld has described how the small team worked in rented offices behind a Texaco gas station, several blocks from Apple's main campus, until its growth forced a move back in early 1982.
Why the moonshots stopped
Alexander Wissner-Gross, a computer scientist and the founder of Reified, came at the problem from history. He noted that one of Teller's grandfathers had invented the hydrogen bomb, and pointed to the era when humanity was landing people on the Moon. Then, he said, the moonshots slowed. Depending on which version of the "Great Stagnation" thesis one accepts, there was perhaps a 50-year stretch, "call it 1970 to 2020-ish," with fewer of them. The term echoes Tyler Cowen's 2011 book The Great Stagnation, which argues that America's easy sources of growth, including powerful technological advances, ran low over the preceding four decades. Around 2020, Wissner-Gross said, moonshots resumed: AI that works, robots that work, and the lipid nanoparticles that helped in the pandemic. He asked whether Teller believed in the pause and how to prevent another one.
Teller first took the dates as a compliment, since X started its "moonshot factory" in 2010. He did think the world had "lost our way a little bit" after the 1960s and the inspiration of NASA, losing what he called "the explorer spirit." Asked for a diagnosis, he said, "I don't know," and then offered a partial answer framed as a "dirty little secret": taking moonshots is "laughably easy."
If you don't care about efficiency, he said, you can find some very energetic, "delusionally optimistic" people, put a lot of money behind them, and you will get some moonshots. It just isn't a good return on investment. During World War II and the Cold War, in his telling, people didn't care about the return. Afterward they probably still wanted moonshots but did care about the return, so they got fewer. X, he said, is trying to make that audacity efficient enough to turn into a system, so that it becomes rational to put much more money in.
Teller said the appetite is there. X is now approached roughly every week by a large company or a country that wants to set up its own moonshot factory. "I'm not sure exactly how we can help them with that," he added, but he expects to see much more of it in the future.
Why not 100 a year
Diamandis pushed on scale. Why does X stop at two moonshots a year? Why not 100 baby moonshot companies a year?
Teller answered that if all he cared about was moonshots and not efficiency, perhaps they would. But efficient innovation depends on a "cultural rhythm" and a set of habits that he called "really easy to describe, but fiendishly hard to get people to practice": playing the long game, being intellectually honest, showing up with humility, and being a team player rather than an egomaniac. Everyone agrees with those values, he said, but he asked the audience to consider honestly whether their own organizations work that way. He heard the laughter. "It's not your fault," he said, but someone would have to be "maniacally determined" to create a microcosm in which those behaviors are the ones that pay off.
That culture engineering is what he focuses on, Teller said, and it is also his limit. "I know how to do that with a few hundred people. I don't know how to do that with a few thousand people."
Wissner-Gross suggested that the ultimate moonshot was therefore the "meta-moonshot" of solving that organizational problem. Teller said the meta moonshot, as X often uses the phrase, is how to systematize radical innovation. That makes the next level a "meta meta moonshot": "how to systematize systematizing radical innovation," or, in effect, how to copy and paste moonshot factories. The world keeps asking X for exactly that, he said, and X would love to help, but he has not yet figured out how to do it.
Diamandis brought in Salim Ismail, the founder of Open ExO, under the banner of his "organizational singularity" idea. Ismail said he might have the methodology. He has spent 10 years building a societal layer, he said, a community now numbering 50,000 people globally, focused on creating environments where new ventures can be spun off very quickly. He borrowed Stewart Brand's "pace layering" model, which describes how fast-moving parts of a society, such as fashion and commerce, generate experiments while slower layers, such as governance and culture, provide continuity (Brand's essay explains the idea). X, Ismail suggested, had solved the inner layers and his community the outer ones, and the two should talk about collaborating.
The investor's clock
The protected space at X still answers to someone. Teller said his investor is Alphabet, and he has to be able to tell Alphabet what he is doing with its money and make it feel good about the flywheel he is creating. That shapes what X works on. He wants the most audacious projects that can move from crazy idea to not-crazy idea in less than a decade, and then to an enduring business in less than another decade. If someone at X proposed going to Mars, he said, his first question would be how that becomes an enduring business within two decades. He added that he wasn't saying it couldn't.
What outsiders underestimate
In the final audience question, Diamandis asked what other companies most underestimate when they try to copy X and fail. Teller came back to culture: "a leadership team that is maniacally focused on engineering a culture" in which people work in the ways that drive radical innovation most efficiently. The hard part, he said, is "making that space, that protected space for those explorers."
He pointed listeners to X's own Moonshot Podcast, whose two seasons feature people who went through the process. Its episodes cover discontinued projects and failed economic tests as well as projects that became businesses. Teller has made the same structural argument in writing. In a 2020 essay, he explained that X was separated from Google because exploratory work needs a different environment from reliable product delivery, and that part of leadership's job is protecting unconventional thinkers.
Ismail argued that the protection itself is the rare part. X, he said, is structured so that "the mothership has left you alone for a long period of time," something he has never seen another company do. Everywhere else, in his experience, the corporate immune system "reaches out its sticky little hands, snatches it back, and then messes it up invariably." He congratulated Teller on lasting this long, and the Google executives who had the discipline to leave X on the edge, "because that's where all the magic happens."