OpenAI hires Noam Shazeer and Dean Ball ahead of IPO

OpenAI has added two high-profile hires as it prepares for a public debut: Noam Shazeer, a foundational engineer in generative AI, and Dean Ball, a former White House AI policy official who wrote he will join OpenAI on July 6 to lead a new Strategic Futures team.

Shazeer announced his departure from Google on Wednesday, where he had worked since 2000 except for a three-year break to co-found Character AI. Two years ago Google re-hired Shazeer in a $2.7 billion deal that gave the company access to Character AI’s technology. Shazeer is credited as a co-author of the 2017 paper Attention Is All You Need, which introduced the Transformer architecture.

Who are Noam Shazeer and Dean Ball?

Noam Shazeer is a long-tenured AI engineer and founder of Character AI, formerly a co-lead at Google’s Gemini effort; Dean Ball is a policy official who briefly worked in the White House and helped publish America’s AI Action Plan. Shazeer spent most of his career at Google, left for three years to start Character AI, then returned under the $2.7 billion agreement two years ago. Ball stepped down from the White House to rejoin the Foundation for American Innovation as a senior fellow before accepting his OpenAI role.

What will Strategic Futures do?

Strategic Futures, which Ball will lead and where he will report directly to Chief Strategy Officer Jason Kwon, will focus on catastrophic risk, recursive self-improvement, labor market impact, and the relationship between frontier labs, governments (particularly the U.S. Federal Government), and society. Ball described the team as "small, high-agency" and said it will cover both public-facing policy and internal governance, adding that internal governance will be central to the future of AI.

How do these hires fit into the broader industry context?

The moves are part of active shuffling among top AI labs including Google, OpenAI, Anthropic, and Meta. Shazeer’s recruitment isolates technical talent across labs; Ball’s hire signals OpenAI is strengthening ties to U.S. policy circles. The article notes a contemporaneous regulatory flashpoint: late last week President Donald Trump ordered an export control ban on Anthropic’s latest models, Fable 5 and Mythos 5, forcing that company to take the models down to avoid noncompliance. The contrast between a rival facing government action and OpenAI adding an insider policy team is explicit in the reporting.

Why it matters

OpenAI’s combination of engineering pedigree and policy experience narrows two persistent gaps for large AI firms: deep technical know-how and direct channels into government decision-making. Shazeer brings work that helped define modern generative models. Ball brings experience drafting national AI policy and a declared mandate to shape frontier AI policy and internal governance. Together they change the mix of talent around OpenAI’s leadership at a sensitive moment for industry regulation.

What to watch

Watch whether Strategic Futures influences regulatory outcomes, and whether Ball’s reporting line to Jason Kwon yields visible shifts in OpenAI’s public policy posture. Also watch for any fallout from Shazeer’s internal controversies at Google: The Information reported he posted opinions on transgender identity and Israel’s war in Gaza that management deleted at Google, and the piece notes it is unclear whether those controversies will follow him. Finally, track competitive impacts from the Anthropic export control action and whether similar government interventions affect model availability.

Sources cited in the reporting include Shazeer’s announcement of his departure, Ball’s July 6 start date and blog post about Strategic Futures, the $2.7 billion agreement that rehired Shazeer two years ago, and the White House export control order affecting Anthropic’s Fable 5 and Mythos 5 models.