Astro Teller, head of Alphabet's "Moonshot Factory" X division, revealed at the TechCrunch Disrupt conference last week that the unit is shifting its approach to bringing ambitious tech projects to market—increasingly opting to spin them off as standalone companies rather than keeping them within Alphabet's corporate structure.

The strategy centers on a dedicated venture capital fund exclusively investing in companies spun out from X, with Alphabet holding only a minority stake. Teller explained, "If Alphabet were the sole limited partner (LP) in this fund, it would still be part of Alphabet’s ecosystem. Even if it invests in X’s projects, they’d remain internal. By keeping our stake small, we avoid undermining our goals."

Named Series X Capital, the fund has raised over $500 million and is managed by former YouTube executive and ex-Facebook CFO Gideon Yu. Unlike Alphabet’s other investment arms—GV (early-stage startups), CapitalG (growth-stage firms), and Gradient Ventures (AI startups)—Series X Capital is legally bound to invest only in X spinouts.

This marks a significant evolution for X, which previously elevated successful projects like Waymo (self-driving) and Wing (drone delivery) into Alphabet subsidiaries. Over the past decade, X realized some moonshots benefit from Alphabet’s resources, while others "could move faster without being tethered to Alphabet, especially when their focus diverges sharply from core businesses."

Teller emphasized the model’s rationale: "Keeping these projects externally linked but uncontrolled allows strategic synergies without constraints." He credited the success of spinouts to X’s culture of "radical honesty," including celebrating the termination of unpromising ideas early.

X defines "moonshots" by three criteria: addressing a global-scale problem, proposing a definitive solution, and leveraging breakthrough technology that offers a "glimmer of hope." Teller noted, "If an idea sounds ‘reasonable,’ it’s not a moonshot by definition."

Projects undergo rigorous testing to identify flaws quickly. "We invest minimally to validate if an idea is crazier—or less crazy—than expected. If it’s the former, we high-five and move on," Teller said. This detachment extends to anonymity—Teller doesn’t know most project originators, including those behind Waymo and Wing—to ensure unbiased evaluation.

The approach yields a 2% "success rate," which Teller reframes as X’s hallmark. Many terminated projects, like AI copywriting tools later absorbed by foundational models, once seemed promising. Spinouts solve a logistical hurdle: previously, X needed external VCs to acquire 51% stakes; now, Series X Capital streamlines the process while maintaining strategic ties.

Employees involved in spinouts receive equity comparable to garage-startup stakes at equivalent funding stages, "with zero personal risk," Teller noted. X recruits by highlighting this trade-off: "Outside startups offer outlier returns, but here, you innovate like a ‘card counter’—no fear, no financial peril."

X staff are paid like other Google employees and receive no equity in early-stage projects—a setup that removes financial pressure to cling to failing ideas. "You can say, ‘This drags us down—kill it,’ without worrying about your kid’s college fund," Teller added.

In 2025, X spun out at least two companies: Taara (optical wireless comms) and Heritable Agriculture (AI-driven crop breeding). Past spinouts include Malta (renewable energy storage), Dandelion (geothermal heating), and iyO (AI headphones). Recently announced moonshot Anori aims to streamline construction workflows via AI, targeting an industry responsible for 25% of global waste and CO₂ emissions.

"Built environments—where we live and work—are foundational to Maslow’s hierarchy and GDP. The sector’s impact is undeniable," Teller concluded.