When Amazon quietly brought in the founding team of Adept last year, it wasn’t a conventional startup acquisition. Instead, it was one of the earliest high-profile examples of what’s now being called a reverse acquihire a hybrid deal where a tech giant hires the key talent and licenses their technology, without actually buying the company.
Adept’s co-founder and former CEO, David Luan, is now leading Amazon’s newly formed AGI Lab the division tasked with chasing artificial general intelligence. In a recent conversation with The Verge, Luan opened up about both the broader AGI mission and the unusual structure of the deal that brought him into Amazon.
Asked about the reverse acquihire trend, Luan was quick to brush off the idea that he wants to be remembered for financial creativity. “I’d much rather be known as an AI research innovator than a deal structure innovator,” he said, while also acknowledging that the move made strategic sense for Amazon. “It’s perfectly rational for companies right now to pull together critical mass in both talent and compute if they want to stay competitive in this race.”
Leaving Adept Behind For a Reason
Luan’s decision to leave Adept wasn’t about abandoning ship; it was about ambition. While Adept had been building useful AI models, he wasn’t interested in pivoting it into a conventional enterprise software company.
“I didn’t want to turn Adept into an enterprise company that only sells small models,” he explained. Instead, he saw an opportunity to work on what he calls the “four crucial remaining research problems” standing between today’s AI and true AGI.
And tackling those problems isn’t something a scrappy startup can do on a budget. “Every single one of them is going to require two-digit billion-dollar clusters to go run it,” Luan said bluntly. “How else am I going to have the opportunity to go do that?”
The Bigger Picture: Talent, Compute, and AGI
For Amazon, the move underscores a broader shift in how major tech players are thinking about the AI race. Traditional acquisitions are giving way to more flexible arrangements that allow companies to bring in top researchers and critical IP without fully absorbing a startup especially when the startup’s long-term goals don’t align with a narrow commercial roadmap.
For Luan, it’s about having the platform to chase AGI at scale. “If you want to solve AGI, you can’t do it halfway,” he implied. That means more talent, more compute, and more patience all things that Amazon, with its vast infrastructure and deep pockets, can provide.
Whether reverse acquihires become a dominant trend remains to be seen. But for now, Amazon’s bet on Luan and his bet on Amazon looks like a calculated move to put serious weight behind AGI research, even if it meant rewriting the usual playbook.
