Why bother with artificial general intelligence (AGI)?

Why not just go straight to superintelligence (ASI)?

That’s precisely the message Meta CEO Mark Zuckerberg is sending with his latest organizational changes made in the last couple of days.

When most of us think of Meta, we think of it as a social media company. After all, Meta owns Facebook, Instagram, and WhatsApp.

But Meta is actually a tech-centric advertising company. It derives almost 99% of its revenues from selling advertising space on its application platforms.

This is what turned Meta into a $1.78 trillion company sitting on $70 billion in cash and generating tens of billions in free cash flow every year. This year, Meta will generate $34.2 billion in free cash flow, which is actually significantly lower than last year’s $52 billion.

And there’s one reason for that… Its ambitions are far greater than advertising.

Skipping Over AGI

Meta is spending at record levels on both data center infrastructure and software development for artificial intelligence.

That’s what has directly impacted its free cash flows. And despite spending more than $100 billion on developing AI, Meta has been significantly behind companies like Google, Anthropic, and xAI in the race to artificial general intelligence.

For one of the largest, cash-rich, and most powerful tech companies, this has been a sore point for Meta that left analysts wondering what Meta might do to catch up.

Now we know.

A couple of weeks ago, Meta absolutely aped into a deal, spending $14.3 billion to acquire only 49% of Scale AI – a high-growth AI service company that focuses on data annotation and AI training data for companies looking to build AI models.

It’s an incredible figure that implies a $28.6 billion valuation for Scale AI – a company that was worth just $1 billion back in 2019.

But it was the deal structure that was telling.

Meta knew it couldn’t get away with an outright acquisition of Scale AI. Scale AI was a neutral party and a critical service provider to the industry. An all-out acquisition would have almost certainly caused major antitrust issues. I doubt the deal would have gone through.

So Zuckerberg acquired 49% to avoid antitrust problems. But the key part of the deal was that Scale AI’s CEO Alexandr Wang will leave Scale AI and join Meta as its Chief AI Officer.

Then, just a couple days ago, we learned that Zuckerberg has established a new division of Meta now known as Meta Superintelligence Labs (MSL) to be led by Wang.

The name says it all…

Meta knows it has fallen behind in the race to AGI. So it’s stepping up to put a team together to develop the next frontier AI model, something beyond AGI.

Meta is gearing up to build an artificial superintelligence.

But don’t take my word for it, here it is directly from Zuckerberg:

As the pace of AI progress accelerates, developing superintelligence is coming into sight. I believe this will be the beginning of a new era for humanity, and I am fully committed to doing what it takes for Meta to lead the way.

And he was clear about why he believes Meta can do it:

Meta is uniquely positioned to deliver superintelligence to the world. We have a strong business that supports building out significantly more compute than smaller labs. We have deeper experience building and growing products that reach billions of people.

And there is no question that Meta is pulling out all the stops and spending whatever it takes to put together a team that might be able to pull it off.

In addition to Zuckerberg’s “acquihire” of Alexandr Wang for $14.3 billion, he also pulled in Nat Friedman from Microsoft, who himself was “acquihired” by Microsoft for $7.5 billion in 2018 when Microsoft acquired GitHub.

Friedman will be helping to lead the new Meta Superintelligence Labs with Wang.

And that’s just two names. Here’s who else is joining Meta to hopefully achieve its ASI ambitions:

• Seven key software engineers who have created various parts of OpenAI’s frontier models.
• A key engineer at Anthropic.
• Three key players from Google, two of whom come from Google’s DeepMind subsidiary.

And I’m sure there are far more in the works.

But what comes next might be hard to believe.

All in on ASI

For the very top talent, Meta is offering packages as high as $300 million over four years and more than $100 million in payouts in the first year. Other team members, like those I listed above, are rumored to be receiving packages around $10 million a year (base, bonus, and equity combined).

Never in my decades in high tech have I ever seen executive recruiting at this level.

Anyone who has actually built a frontier model, or any leading AI company, is invaluable in this environment. And these kinds of comp packages are only possible because a handful of companies out there have so much capital available to them that they can afford these astronomical packages.

But let’s not let these extraordinary compensation packages distract us from the far bigger story.

“Developing superintelligence is coming into sight.”

Many in the industry are starting to believe that artificial general intelligence is now becoming a near-term event. This is a prediction that I’ve made for quite some time. We’re already seeing sprouts of AGI.

Something very big is right around the corner, and I predict that xAI’s Grok 4 – which is expected to be released shortly after July 4 – is going to leap far ahead of the rest of the industry. Many will suggest that Grok 4 is an early version of AGI.

We’re so close…

Which is why Meta isn’t setting its target on such an imminent technology. It is assembling a team to build, and hopefully lead, what’s coming next…

… namely, artificial superintelligence.

Jeff