Loyalty No Longer Exists in Silicon Valley

Since the middle of last year, Silicon Valley has witnessed at least three significant AI “acqui-hires.” Meta committed over $14 billion to Scale AI, acquiring its CEO, Alexandr Wang; Google shelled out $2.4 billion to license Windsurf’s technology, integrating its cofounders and research teams into DeepMind; and Nvidia invested $20 billion in Groq’s inference technology, hiring its CEO and other team members.

Meanwhile, the leading AI labs have been engaged in a high-stakes, seemingly endless game of talent reshuffling. This latest wave began three weeks ago when OpenAI announced it was bringing back several researchers who had left less than two years prior to join Mira Murati’s startup, Thinking Machines. Concurrently, Anthropic, founded by former OpenAI staff, has been attracting talent from the ChatGPT creator. In response, OpenAI just hired a former safety researcher from Anthropic as its “head of preparedness.”

The hiring activity in Silicon Valley illustrates the “great unbundling” of the tech startup, as described by Dave Munichiello, an investor at GV. In the past, tech founders and early employees often remained until either the company closed or there was a significant liquidity event. Today, with generative AI startups rapidly expanding and flush with capital, the focus on research talent means “you invest in a startup knowing it could be broken up,” Munichiello explained.

Founders and researchers at prominent AI startups are moving between companies for various reasons. A primary motivator, naturally, is financial gain. Last year, Meta reportedly offered leading AI researchers compensation packages reaching into the tens or hundreds of millions, giving them not just access to advanced computational resources but also … generational wealth.

However, the motivations aren’t solely financial. Cultural shifts within the tech industry have left some employees hesitant to commit to a single company for extended periods, according to Sayash Kapoor, a computer science researcher at Princeton University and a senior fellow at Mozilla. Employers once confidently expected workers to stay until their stock options vested, typically after four years. In the idealistic landscape of the 2000s and 2010s, many early cofounders and employees genuinely believed in their companies’ missions and sought to contribute to them.

Now, Kapoor notes, “people recognize the limitations of the institutions they work for, and founders approach things more pragmatically.” For instance, the founders of Windsurf might believe their impact could be greater at a resource-rich company like Google, Kapoor suggests. He also points out that a similar trend is occurring in academia, with more PhD researchers exiting their computer-science programs for industry roles over the past five years. Staying put carries higher opportunity costs in an era of rapid AI innovation, he mentions.

Investors, wary of getting caught in the crossfire of the AI talent wars, are taking steps to safeguard their interests. Max Gazor, the founder of Striker Venture Partners, explains that his team is prioritizing “chemistry and cohesion” when evaluating founding teams more than ever before. He notes that it’s becoming increasingly customary for agreements to include “protective provisions requiring board approval for significant IP licensing or similar situations.”

Gazor observes that some of the largest recent acqui-hire deals involved startups established well before the current generative AI surge. For instance, Scale AI was founded in 2016, a time when the deal Wang secured with Meta would have been unimaginable to many. Today, however, such potential outcomes might be factored into early term sheets and “constructively managed,” Gazor explains.