Meta looked at buying Perplexity, SSI, and Thinking Machines

It might soon be easier to list the artificial intelligence startups Mark Zuckerberg hasn’t tried to snap up.

Meta held recent acquisition talks with Ilya Sutskever’s Safe Superintelligence (SSI), former OpenAI CTO Mira Murati’s new venture Thinking Machines Lab, and AI search challenger Perplexity, according to reporting from The Verge. None of the discussions progressed to a formal offer due to disagreements over price and strategic vision, but the aggressive push shows just how urgently Zuckerberg wants to reboot Meta's AI efforts.

Building a multi-billion-dollar dream team

While outright acquisitions stalled, Zuckerberg’s talent blitz moved ahead at full throttle. Meta just paid more than $14 billion to quickly hire former Scale AI CEO Alexandr Wang. Wang said goodbye to his team last Friday and was already working inside Meta by Monday, rounding up talent for a brand-new AI division.

Key pieces of that structure are emerging. SSI co-founder Daniel Gross and former GitHub CEO Nat Friedman are poised to co-lead the Meta AI assistant group, reporting directly to Wang. Meanwhile, the startup founders who walked away from Meta's overtures are raising massive capital elsewhere. Sutskever and Murati each recently pulled in a couple of billion dollars despite neither being close to shipping a product, while Perplexity is finalizing roughly $500 million.

Industry insiders note that nine-figure and even ten-figure compensation packages are no longer shocking for premier AI researchers, especially given how high valuations have climbed at rivals. OpenAI CEO Sam Altman appeared somewhat defensive this week, telling a podcast that none of his company's best people were leaving for Meta and questioning whether huge upfront payouts foster good culture—though critics quickly noted OpenAI itself recently shelled out $6.5 billion to bring in Jony Ive and his hardware startup.

Betting on frames over screens

All this high-level talent is pointing directly toward Meta's hardware roadmap. Alex Himel, Meta’s VP of wearables, met with Wang this week to focus on tailoring AI models specifically for eyewear.

Himel explained that Meta’s philosophy on scaling wearables has shifted dramatically. While the company once thought mainstream adoption required bulky AR displays like the Orion prototype, leadership now believes AI functionality can drive mass adoption much faster. Audio and media capture remain the top two use cases on current frames, but AI usage is climbing steadily. Meta is already running 1-billion-parameter models directly on the glasses frame to handle egocentric video and audio feeds without cloud latency, and it is expanding its EssilorLuxottica partnership from Ray-Ban into performance-focused Oakley designs.

Why it matters

Zuckerberg is effectively attempting to purchase an elite AI lab whole-cloth after missing out on early generational breakthroughs, and he is willing to hand over enormous leverage to do it.

For regular consumers, the clearest payoff will arrive on your face rather than inside a browser tab. If Meta can successfully blend on-device small models with lightweight eyewear, ambient computing becomes a practical daily tool instead of an awkward sci-fi demo. Whether you are walking around or working out, an assistant that sees what you see without constant cloud lag represents genuine utility.

For the tech industry, Meta’s willingness to throw 10-figure checks at individual leaders sets a brutal precedent. It forces incumbents like OpenAI to play intense internal defense just to retain core talent.

We think the hardware-first focus is the smartest part of this pivot, but assembling an all-star roster of former CEOs creates immediate friction. Managing the egos and overlapping mandates of Wang, Gross, and Friedman under Meta's legacy corporate structure looks like an organizational minefield. If this new leadership tier clashes over strategy, Zuckerberg will have spent billions simply buying executive turbulence.