Two hires in one week tell you more about OpenAI's current priorities than any press release would.
Noam Shazeer is not a typical engineering executive. He co-authored the 2017 research paper that introduced the design principle now used by every major AI system on the market, including the one that powers ChatGPT. Google valued his knowledge so highly that it paid $2.7 billion in 2024 to bring him back after he left to start a chatbot company. He had been co-leading Google's Gemini AI models. Now he is joining OpenAI.
His role at OpenAI has not been publicly defined yet, and neither company offered detail beyond confirming the hire. But the move signals that OpenAI wants someone at the very top of the technical tree, not just a good manager, sitting inside its research operation.
The second hire is Dean Ball, who starts July 6 as head of a new unit called Strategic Futures. Ball previously served as a senior policy adviser for AI at the White House Office of Science and Technology Policy, where he helped write the current administration's AI Action Plan. His new team will report to OpenAI's chief strategy officer and will focus on government relations, internal governance, and what Ball describes as catastrophic risk and long-term societal questions.
The governance piece is worth noting. Ball has written publicly that AI companies will, almost unavoidably, have to make major decisions about how AI is built and deployed before governments catch up with legislation. He wants to be inside the room where those calls happen.
Both hires make more sense when you look at the backdrop. OpenAI has confidentially filed paperwork for a public stock market listing, targeting a valuation somewhere between $850 billion and $1 trillion. The company is generating around $2 billion in revenue per month, but is not yet profitable and does not expect to be until roughly 2030. Before it can list, it also needs to complete a restructuring from a non-profit to a standard for-profit company.
At the same time, Anthropic, OpenAI's closest competitor, has been in direct conflict with the U.S. government. The Trump administration ordered Anthropic to pull its two most powerful models offline, citing national security concerns after reports that someone had found a way to bypass the models' safety limits. Anthropic had to shut those models down for all customers, globally, to comply.
OpenAI has so far avoided that kind of direct government confrontation. Bringing in someone with Ball's connections and policy background is a deliberate move to keep that advantage intact.
For anyone in a business that uses or is considering AI tools: the larger pattern here matters. Governments are starting to treat advanced AI models the way they treat weapons technology, controlling who can access them and under what conditions. Anthropic's situation shows this can happen fast, with little warning, and affect paying customers immediately. Which AI provider you depend on is no longer a purely technical or commercial question. It is also a question of how well that provider manages its relationship with the governments that can flip a switch and take it offline.