Meta hires two more researchers from OpenAI: a new wave of record-breaking hires in the race for super AI models

Meta recruits Allan Jabri and Lu Liu, key figures in multimodal AI

On July 11, 2025, the AI ecosystem was shaken by significant news: Meta has hired Allan Jabri and Lu Liu, two leading OpenAI researchers specializing in the multimodal frontier, thereby strengthening its flagship laboratory, Superintelligence Labs. The announcement, reported by The Information, confirms that Mark Zuckerberg’s company is stepping up its efforts to secure the best minds in artificial intelligence, with the goal of competing head-to-head with OpenAI, Google DeepMind, and the other giants of the new AI era.

The Superintelligence Labs strategy: an unprecedented hiring campaign targeting former OpenAI employees

Jabri and Liu’s case is not an exception, but the latest stage of an aggressive strategy that in just a few weeks has already brought three other former OpenAI researchers—Lucas Beyer, Alexander Kolesnikov, Xiaohua Zhai—to Meta, along with prominent names from Apple, DeepMind, and Anthropic. The approach is clear: Meta offers multimillion-dollar compensation, stock options, and full research autonomy to attract and retain the most sought-after talent in the sector. The new Superintelligence Labs team is built to develop generative models capable of processing and producing text, images, audio, and video, aiming to compete directly with GPT-4o, Gemini Advanced, and Claude Opus.

Behind the scenes of Meta’s colossal offers: packages up to $100 million (and $25 million upfront), personal calls from Zuckerberg, and OpenAI’s counter-retention program

In recent months, Meta has ramped up one of the most aggressive recruiting campaigns Silicon Valley has ever seen to lure top researchers away from OpenAI. According to a leak reported by The Information and confirmed by several internal sources, the Menlo Park company has put forward compensation packages that can reach up to $100 million—a figure that includes stock units distributed over four years, performance bonuses, and, for a few strategic leaders, upfront cash. Meanwhile, recruiting emails circulating on Slack indicate that the average upfront signing bonus for senior researchers is still around $25–30 million in immediately vested RSUs, with substantial annual stock refreshes on top.

The persuasion strategy goes beyond mere financial incentives: in several cases, the offers have been accompanied by personal calls from Mark Zuckerberg, who explained Meta’s vision of “open superintelligence” and promised privileged access to H100 clusters and upcoming MTIA-3 chips, along with a “no-limits” research environment featuring independent budgets for academic publications. An internal OpenAI memo—leaked to the press—describes these proposals as “financially eye-watering,” but warns about the risk of three-year lock-in clauses.

OpenAI’s response was swift: the board authorized a $500 million retention program that includes retroactive raises, accelerated stock refreshes, and an unprecedented six-month paid “research sabbatical” for senior researchers. Nonetheless, Meta seems determined to continue its assault, convinced that a mix of economic resources, visibility, and research freedom can tip the scales in the global race for super AI.

Competition over working conditions and new retention policies in the AI sector

One of the less visible but fundamental aspects of this phase concerns how Meta, as well as OpenAI, DeepMind, and other big tech companies, are rethinking the very concept of a work environment to attract and retain the most valuable talent. It is no longer just about offering stellar salaries, but also ensuring research freedom without constraints, direct access to top-tier hardware infrastructures, and the ability to publish innovative academic work without restrictions. Some former OpenAI researchers have highlighted how, in career choices, the ability of companies to offer a context where ideas can be experimented with quickly, without the bureaucratic delays typical of traditional multinationals, is becoming increasingly important.

The growing weight of acquisitions and collaborations between startups and industry giants

Another key dynamic concerns the increasing interconnectedness between emerging AI startups and major players like Meta, Google, Amazon, and Microsoft. On the one hand, these acquisitions and hiring campaigns can accelerate the diffusion of new technologies and facilitate know-how transfer. On the other, they risk increasing the concentration of skills and resources in just a few hands. The current trend shows ever-closer collaboration between companies that, even while directly competing in the market, do not hesitate to strike deals for sharing datasets, algorithms, and development tools, all to accelerate sector evolution and remain globally competitive.

Ethical and regulatory challenges: transparency, governance, and the need for new rules

The rapid movement of top researchers from one company to another also raises major ethical and regulatory questions. Antitrust authorities in the United States and Europe are beginning to closely monitor these dynamics, questioning how the control of human and intellectual resources can affect competition and the market’s ability to innovate healthily and openly. At the same time, there is growing pressure to introduce shared codes of conduct, transparent governance criteria, and greater clarity regarding the responsibilities of AI developers, especially as these talents work on increasingly powerful models that could have a major impact on society.

________

FOLLOW US FOR MORE.

DATA STUDIOS

datastudios.org