Meta Paid this Guy a Record $200m To Leave Apple

Meta Platforms has just written what may be the fattest signing cheque in Silicon‑Valley history. Bloomberg reports that the company wooed Ruoming Pang, the engineer who led Apple’s foundation‑model team, with a pay package worth more than US $200 million spread over several years—figures that dwarf even Wall‑Street banker bonuses. Pang joins Meta’s fast‑growing Superintelligence Lab, overseen by former Scale AI chief Alexandr Wang and newly hired venture technologists Nat Friedman and Daniel Gross.

Pang is only the latest blue‑chip scalp in what insiders call Mark Zuckerberg’s “talent land‑grab.” Since April, Meta has lured at least a dozen senior researchers from OpenAI, Google DeepMind, Anthropic and now Apple, with some offers topping US $100 million in first‑year compensation, according to multiple press accounts and industry recruiters. The company also invested US $15 billion in Scale AI last month, effectively buying both data‑labelling capacity and Wang’s leadership in one stroke.

Why such eye‑watering sums you may ask? Zuckerberg is convinced that next‑generation AI—what he calls “superintelligence”—will be won by a handful of breakthrough algorithms that require a small cohort of world‑class architects, not armies of generalist coders. If securing those architects costs a few hundred million dollars apiece, the argument goes, the payoff in product leadership (and a trillion‑dollar market cap bump) will more than compensate.

Early signs suggest the approach is moving the needle. Meta insiders say internal preview builds of Llama 5 are “substantially ahead” of the public Llama 4 release that disappointed many users in May. Meanwhile, Apple’s own AI roadmap has reportedly slipped twice in six months, and the loss of Pang could widen that gap.

Yet the strategy is not without risk. Offering nine‑figure packages resets salary baselines across the industry, making it harder for Meta to retain existing staff and raising questions about sustainability if the broader economy cools. Executives at OpenAI and Anthropic have publicly wondered whether a collection of star hires can cohere into a durable research culture—or whether the company will simply become a “Galácticos” squad whose egos outstrip its execution. There is also the matter of shareholder patience: Meta’s AI build‑out already commands a projected US $45 billion in 2025 capex, and returns may take years to materialise.

In the short term, though, the Pang coup underscores a brutal reality for Big Tech: the race to hire the handful of people who can bend large‑language‑model performance curves is escalating faster than any chip shortage. Whether Zuckerberg’s high‑stakes gambit becomes a masterstroke or an expensive footnote will hinge on one question—can Meta’s newly assembled brain‑trust translate record‑setting pay packets into breakthroughs that ordinary users actually feel?