[News] Meta Reshapes AI Division in Superintelligence Push Amid Talent War and Heavy Spending

According to Bloomberg, Meta is breaking up its newly created AI division, Meta Superintelligence, into four separate teams and reallocating much of its existing AI staff in a bid to better leverage billions of dollars’ worth of recently acquired talent.

As Bloomberg notes, citing The Information, Meta Superintelligence Labs is being reorganized into four units: TBD Lab, headed by Alexandr Wang, which will oversee the company’s large language models; FAIR, Meta’s long-standing AI research arm; Products and Applied Research, led by former GitHub CEO Nat Friedman; and MSL Infra, dedicated to the costly infrastructure required to power Meta’s AI ambitions.

Citing an internal memo from Tuesday by Alexandr Wang—the former Scale AI CEO who recently became Meta’s chief AI officer—the report says the reorganization is intended to speed up the company’s push toward what it calls superintelligence.

Restructuring and Strategy

Sources said Tuesday’s reorganization did not include layoffs, Bloomberg notes. Yet Meta’s AI leadership has seen repeated shake-ups in recent years, including several this year as the company scrambles to keep pace with OpenAI and Google, according to Bloomberg.

In addition, Meta is weighing an overall downsizing of its AI division—which may involve cutting roles or reallocating staff to other parts of the company—after the unit swelled to thousands of employees in recent years, according to sources cited by The New York Times.

Beyond the restructuring, The New York Times, citing sources, reports that Meta is actively considering the use of third-party AI models. This could involve building on freely available open-source models or licensing closed-source models from other companies, the report notes.

Spending and the Talent Race

Mark Zuckerberg, Meta’s chief executive, has been pushing aggressively into AI. To lure talent from rivals like OpenAI and Google, the company has dangled nine-figure pay packages—fueling a fierce poaching battle in Silicon Valley, according to The New York Times.

To speed up its AI push, the company last month raised the lower end of its annual capital spending forecast by $2 billion, setting a new range of $66 billion to $72 billion, as  Reuters notes.

Read more

• [News] Meta’s Poaching Effort at Ex-OpenAI CTO’s Startup Fails as AI Talent War Escalates
• [News] Meta Reportedly Teams Up with Broadcom, Taps Quanta for Next-Gen ASIC-powered AI servers

(Photo credit: Meta)