Alphabet’s X moonshot factory is changing its approach to launching bold tech ventures, now favoring spinning them out as independent startups rather than keeping them under Alphabet’s umbrella, according to X chief Astro Teller at TechCrunch Disrupt last week.
This new method relies on a specialized venture fund created solely to back X spinouts, with Alphabet holding only a minority stake. “If Alphabet were the only limited partner, the fund would be internal, so any investment in an X project would still be within Alphabet,” Teller said on stage. “Alphabet can be a small LP, but if it’s more than that, we defeat the purpose of what we’re trying to do.”
That fund, Series X Capital, has secured over $500 million and is managed by Gideon Yu, a former executive at YouTube and ex-CFO of Facebook. Bloomberg first disclosed the fund’s existence last year. Unlike Alphabet’s other investment branches—GV, which targets early-stage startups; CapitalG, which focuses on growth-stage companies; and Gradient Ventures, which invests in AI—Series X Capital is legally required to invest only in companies spun out from X.
This marks a significant shift for X, which previously transitioned successful projects like Waymo and Wing into Alphabet subsidiaries. Teller noted that over the past ten years, the team realized some moonshots thrive with Alphabet’s backing, but others “move faster and don’t really gain from being part of Alphabet because they’re fundamentally different.”
“Placing them just outside Alphabet’s boundaries allows us to stay closely connected and gain strategic advantages, but without direct control,” Teller said.
At Disrupt, he emphasized that this spinout model is possible only because of X’s commitment to intellectual honesty, including a culture that openly rewards shutting down even promising ideas.
X defines a moonshot as having three elements: it must tackle a massive global issue, propose a product or service that could eliminate that problem, and use breakthrough technology that gives the team a “glimmer of hope” they can solve it. Importantly, Teller said, “if a proposed moonshot sounds reasonable, we’re not interested, because by definition, that’s not a moonshot.”
What happens to ideas that fit these criteria? X rigorously tests them, actively seeking reasons to abandon them, Teller said. “If you suggest something that sounds pretty wild, meets those three standards, and is a testable hypothesis, we can spend a small amount to find out if it’s even crazier or less crazy than we thought,” he explained. “If it turns out to be even crazier, great, high five, let’s kill it and move on.”
This process requires people to separate themselves from their ideas, which is why Teller said he often doesn’t know who originated most X projects, including Waymo and Wing, which now delivers Walmart packages in about six U.S. cities. “If you’re leading a project and see it as ‘your baby,’ how likely are you to be truly intellectually honest?” he asked the Disrupt audience.
In reality, this means X tackles the most challenging aspects of projects first, intentionally searching for reasons to end them. The result is a tough 2% success rate, which Teller sees not as a flaw but as a feature. X has ended far more projects than it has launched, including entire categories that once looked promising, such as copywriting AI tools that were eventually overtaken by foundation models.
All this experimentation and failure comes at a cost. The spinout model addresses a practical issue: previously, X needed to find outside investors willing to take at least 51% ownership to spin a project out of Alphabet. Now, with a fund that “truly understands us” and is “legally bound to invest only in our spinouts,” Teller said, X can make the spinout process more systematic while keeping strong strategic connections.
Even though X encourages detachment from ideas, employees still have a significant stake when projects become independent. For those working on spinouts, the financial rewards are considerable. “You and your team will own a portion of the new company,” Teller said. “It’s about the same as if you’d started it in your garage at that funding stage, but without taking on any personal risk.”
Potential X recruits are told about this trade-off up front. “Your chance for a huge windfall is greater outside, I admit that,” Teller said at Disrupt. “But if you join X, you get to innovate with us, counting cards in the game of invention, with no fear and no personal financial risk.”
X staff are compensated like other Google employees and don’t get equity in early-stage projects, since “it’s not even a company yet; it’s just an idea we’re exploring,” Teller said. This removes the financial pressure that might stop founders from abandoning their own projects. “You can say, ‘This one isn’t improving our average, let’s toss it,’” Teller said. “And since you haven’t risked your kids’ college fund, it’s not scary.”
In 2025, X has spun out at least two companies: Taara, which works on wireless optical communication, and Heritable Agriculture, a biotech startup using machine learning to speed up crop breeding. Earlier spinouts that attracted outside investment include Malta (renewable energy storage), Dandelion (geothermal heating), and iyO (AI-powered earbuds).
Right before Disrupt, X introduced its latest moonshot: Anori, described as “a new AI platform designed to help real estate developers, architecture and construction firms, and cities navigate the complexities of new building projects.” When asked what makes this AI platform a moonshot, Teller cited the scale of the challenge—and the opportunity.
“The built environment accounts for about 25% of the world’s solid waste and roughly a quarter of global carbon dioxide emissions. It’s fundamental to human needs—it’s where we live and spend most of our time. It also represents a significant share of global GDP. So, it’s hard to find an industry that matters more.”
You can watch our full conversation with Teller here, starting at the 6:08 mark.
