Alphabet's X moonshot factory is changing the way it brings ambitious technology projects to market, with projects increasingly being spun out as independent companies rather than staying within Alphabet's corporate structure, X head Astro Teller revealed on TechCrunch Disrupt last week.
The strategy relies on a dedicated venture fund that exists solely to invest in the spinout of Company X, with Alphabet being only a minority investor. “If Alphabet were the only LP, the fund would be in Alphabet, and when you invest in something in Company X, it would still be in Alphabet,” Teller explained on stage. “So Alphabet can be a small LP, but if it becomes more than a small LP, it will undo what we are trying to accomplish.”
That fund is Series X Capital, which has raised more than $500 million to date and is run by former YouTube executive and Facebook CFO Gideon Yu. Bloomberg first reported on the fund's existence last year. Unlike Alphabet's other investment arm, GV, it invests broadly in early-stage startups. CapitalG supports growth stage companies. Gradient Ventures — Series X Capital, which invests in AI startups, is legally required to invest only in companies spun out from X.
This approach represents a meaningful evolution for X, which has historically graduated successful projects like Waymo and Wing to independent subsidiaries of Alphabet. Teller said the lab has learned over the past decade that while some moonshots benefit from Alphabet's resources and scale, other projects “can move faster and are so different that they don't really benefit from being part of Alphabet.”
“Landing it just outside of Alphabet makes sense because we can work closely with them and have a lot of strategic co-benefits with them, but we don't necessarily have control over them,” he said.
At Disrupt, Teller explained that the spinout strategy works because of Company X's ruthless approach to intellectual integrity, including a culture that actively celebrates killing promising ideas.
X defines that a moonshot has three specific elements. That means trying to solve a huge problem in the world, proposing some kind of product or service that solves that problem, and leveraging breakthrough technology that creates a “glimmer of hope'' that a team within X can solve that problem. Critically, Teller said, “If someone is proposing a moonshot and it seems reasonable, the company is not interested, because by definition it is not a moonshot.”
What happens to ideas that meet these criteria? X is ruthlessly testing them, looking for reasons to kill them, Teller said. “If we propose something and it sounds pretty wild, and it has those three elements and it’s a testable hypothesis, then if we pay a small amount of money, we can learn something about whether it’s a little bit crazier than we thought, or a little less crazy than we thought,” Teller explained. “If it's a little crazier than we thought, cool, high five, let's put a bullet in its head and move on.”
tech crunch event
San Francisco | October 13-15, 2026
Because this approach requires separating people from their ideas, Teller said he doesn't even know who started most of Company X's projects, including self-driving car company Waymo and drone delivery company Wing, which currently delivers packages for Walmart in about six U.S. cities. “If we're going to go explore something, and you [as the lead inventor] If you're feeling like, “This is my baby,” what chance do I have of getting you to practice true intellectual honesty? ” he told the Disrupt audience.
In practice, this means that X tackles the most difficult parts of the project first and is actively looking for reasons to stop the project. The result is a brutal 2% hit rate, which Teller sees as a feature rather than a failure. X has canceled far more projects than he has ever started. That includes entire categories that were once seen as promising, such as copywriting AI tools, which Foundation Model eventually absorbed.
Testing and failure can be costly. Spin-out structures solve practical problems. Previously, X had to find an outside venture investor willing to take over at least 51% of the business to spin out of Alphabet, but by creating a fund that “understands us deeply” and is “legally obligated to invest only in what comes out of us,” Teller said X can structure the spinout process while maintaining close strategic relationships.
Despite the emphasis on distancing themselves from the idea, X's employees actually play an important role when the project spins out. For those working on projects aimed at independence, the financial incentives are substantial. “You and the rest of your team will get a piece of that company,” Teller said. “At that stage of financing, it’s about the same amount of money you would get if you started out of your garage without taking any risks during that time.”
X's pitch to potential employees also clearly explains this trade-off. “The increase of four or five standard deviations is going to be even bigger on the outside, I admit that,” Teller said on Disrupt. “But if you come to Company X, you can become an innovation card counter with us without fear and without taking any financial risks.”
Company X's employees are paid like other Google employees, but there is no capital for early-stage projects. That's because “Company X isn't even a company; it's an idea that we're trying to learn from,” Teller explained. This removes the financial pressure that prevents founders from killing their ideas. “You can say, 'Hey, this isn't going to improve your average, let's throw this out,'” Teller explained. “And you're not risking your child's college funds, so it's not scary.”
X is spinning out at least two companies in 2025. One is Taara, which develops wireless optical communication technology, and the other is Heritable Agriculture, a biotechnology company that uses machine learning to improve crop breeding. Previous spinouts that have raised external funding include Malta (renewable energy storage), Dandelion (geothermal heating), and iyO (AI-powered earphones).
On the eve of Disrupt, X announced its newest moonshot company, Anori. It is, as the company describes it, “a new AI platform that helps real estate developers, the A&C industry, and cities untangle the complexities of new building projects.” Asked on stage why this particular AI platform is a “moonshot,” Teller pointed to the scale of the problem and the opportunity.
“The built environment accounts for approximately 25% of the world's solid waste. [and] Approximately 25% of the world [carbon dioxide] output. It is literally based on Maslow's hierarchy of needs and is where we live and where we spend most of our time. This accounts for a large portion of the world's GDP output. So it would be hard for us to become more important as an industry. ”
You can watch the entire conversation with Teller starting at the 6:08 mark.