AI Talent Values Freedom Over Pay, Experts Say

Two years after Google paid over $2 billion to acquire Noam Shazeer and part of his CharacterAI team, Shazeer departed for OpenAI, per Axios (June 18, 2026). Shazeer co-authored the 2017 paper "Attention Is All You Need," which introduced the Transformer architecture underlying virtually all modern large language models. Within days, John Jumper - who shared the 2024 Nobel Prize in Chemistry for his AlphaFold work at DeepMind - announced he was leaving Google DeepMind for Anthropic, per Axios (June 23, 2026). In the same period, Nvidia acqui-hired the Essential AI team including Ashish Vaswani, another co-author of "Attention Is All You Need." Axios reports that Google DeepMind lost two high-profile researchers in a single week, in a period marked by broader lab-to-lab defections.

Axios reporting identifies a multi-factor decision calculus for top researchers: potential financial returns including IPO upside (OpenAI and Anthropic are both approaching public market events), access to compute, each lab's perceived position in the AGI race, and alignment with leadership values. Business Insider, citing VC Jason Lemkin on the 20VC podcast, adds a non-financial dimension: the ability "to work on the problems they care about with fewer constraints." These factors together - not salary alone - explain why billion-dollar acqui-hire arrangements can fail once lock-up periods expire.

For engineering leads and hiring managers, the week illustrates two concrete dynamics. First, acqui-hire retention is time-bounded: the Shazeer deal secured roughly two years. Second, the perceived winner-takes-most framing in the AGI race is now a pull factor independent of cash compensation - researchers are reported to be selecting the lab they believe will define the next era of AI development. Forthcoming IPOs at OpenAI and Anthropic offer liquidity that already-public incumbents cannot easily match. Axios also notes a compounding risk: top researchers recruit other sought-after scientists, so a single high-profile exit elevates the risk of further defections.

Google DeepMind CEO Demis Hassabis publicly maintained confidence in the company's ability to attract and retain talent, per Semafor (June 23, 2026). Signals to track: whether Google responds with structural changes to equity or research governance, and whether OpenAI and Anthropic IPO timelines stay on track - which will determine whether the financial pull factor for frontier talent strengthens or dissipates.