If you enjoy this article, consider sharing it.

Share

Since the Spring we’ve known that OpenAI is working on a social network that is likely to be a Twitter-clone, that could compete directly with Meta’s Threads product. While Mark Zuckerberg’s dream of a Metaverse didn’t exactly materialize in short order, Facebook has always believed the best defense, is offense.

In CNBC’s 2025 edition of its Disruptor 50 series, AI and National defense companies feature heavily. While we know that Meta is collaborating with Anduril on developing AI-powered virtual and augmented reality devices for the U.S. military, Meta is technically rich enough to compete for Superintelligence glory and the AGI race. Well, that’s exactly what it is doing we found out this week.

I cannot believe I’m saying this, but:

Meta launching AI superintelligence lab with nine-figure pay push, reports say

Meta is preparing to launch a new artificial intelligence research lab dedicated to the pursuit of "superintelligence" -- a hypothetical AI system where Mark Zuckerberg is personally leading this project.

As part of the deal, Meta will take a 49% stake in the data-labelling and annotation startup, The Information reported. This implies a doubling of Scale AI’s valuation.

Meta has failed in recent months in its latest iteration of Llama 4, to keep up with China in open-source models and there has been a reported exodus of AI talent at its company. It’s not clear how Meta is therefore capable of entering, or even winning, the AGI race. They have the funds to do it, but not necessarily the AI talent.

Zuckerberg while insisting on efficiency at Facebook, is known for splashing money into R&D and some fairly odd ideas about the future of tech. Meta’s rumored huge deal investing around $14 Billion into Scale AI (Meta was already an investor) is intimately part of this project. Meta has invited Alexandr Wang, the 28-year-old Chinese-American Billionaire founder and CEO of Scale AI , to join the new lab. As part of the deal, Meta is in talks to invest billions of dollars in Scale AI, which will also bring other Scale AI employees to Meta.

Zuck Going AGI Mode

According to Bloomberg reporting, Zuckerberg has been personally leading recruiting for the team of about 50 people — and has even rearranged Meta's offices so the new hires sit near him.

This team is part of a new AI research lab focused on creating the world's most advanced AI platform, aiming to outpace competitors like OpenAI and Google in the race for AI dominance.

All things considered Anthropic, Google, Alibaba, OpenAI, Thinking Machines Lab and many others may already have a leg up on Meta, including DeepSeek. There’s not much evidence Meta can even make decent LLMs that are bleeding edge as of 2025 inspite of investing Billions in AI infrastructure and cutting-edge datacenters.

Mark Zuckerberg sounds like a frustrated tycoon for being so wrong about his directions around the future of technology. Axios reports that both the Times and Bloomberg reported that Zuckerberg has been personally frustrated with the pace of Meta's AI development and public stumbles, and has gone into a more hands-on mode to accelerate progress.

Not unlike Musk’s recent obsession with xAI and Grok, Meta’s Superintelligence team represents a high-profile, well-funded effort to achieve AGI and superintelligence, aiming to integrate these capabilities across Meta’s products and maintain competitive leadership in AI technology. If you consider Meta’s ties on Military AR glasses with Anduril, this could also be about building AGI for the U.S. military. Scale AI is already deeply entrenched with Palantir and others along those lines of automating more aspects of the DoD and U.S. military with AI.

The AI Talent Wars are heating up

Meta has offered seven- to nine-figure compensation packages to dozens of researchers from leading AI companies such as OpenAI and Google, some of whom have already agreed to join.

Anthropic already leads in this with over an 80% retention rate for employees hired over the last two years. That’s according to Venture Capital firm SignalFire’s recently released 2025 State of Talent Report, which analyzed tech hiring and employment trends.

New Companies are outcompeting Older ones for AI Talent

As more serious AI startups scale and get further ahead as compared with BigTech incumbents (Google excluded), the new AI Behemoths could evolve to challenge the business models of those like Meta, Apple and Google. I initially read reporting that the Scale AI deal with more related to automating digital Advertising so I’m a bit surprised to hear the Superintelligence aspect of all of this.

Is Meta conducting an acquihire of Scale AI? And how would that lead to Superintelligence? It appears Alexandr Wang (with claims to being the youngest self-made Billionaire) got the better deal in this concerning combination of companies.

What is an Acquihire?

An "acquihire" is the acquisition of a company, primarily for the purpose of hiring its employees rather than acquiring its products, services, or other assets. It's a way for companies to quickly onboard a skilled team by buying a small company and integrating its employees.

The reported $14 billion investment in Scale AI is strategic not just for financial reasons but for technical capabilities. What could Alexandr Wang do leading the new lab? It’s a pretty shocking development.

Meta CEO Mark Zuckerberg has invested billions of dollars to turn the company into an AI powerhouse but Meta really stumbled badly with Llama 4. Scale AI provides data labeling services to companies such as Microsoft and OpenAI to help them train their AI models. Meta acquiring them, might actually harm OpenAI and Microsoft’s ability to train their models sometime in the future. Does Mark Zuckerberg think he can actually compete with the likes of Anthropic, OpenAI and Google in frontier LLMs?

Microsoft, Amazon, Google and Meta Desperate to Keep Up

• Amazon's total investment in Anthropic is $8 billion.

• Microsoft’s total investment in OpenAI is at least $13 Billion.

• To see Meta investing (and merging some of) Scale AI into their labs for $14 Billion bizarre because Scale AI isn’t at the same level as those two companies that are the defacto Duopoly of Generative AI.

Talent and Energy are the new Bottlenecks 2025-2030

It’s just not very fiscally responsible for Meta to do this. Meta already over-spends on AI compared to its leverage with Generative AI. Meta's capital expenditures (capex) for 2025 are expected to be in the range of $64 to $72 billion, including principal payments on finance leases. If you add another $10 billion or more, it’s absurd spending.

If you were a top AI research or engineer, would you even want to work at Meta under Zuck or a 28-year old Tycoon? It’s not just about money. Talent wants to work with other top talent and in situations where AI ethics and values match their own preferences. In 2025, that’s not been OpenAI or Meta in case you haven’t noticed:

• AI churn of talent at OpenAI to Anthropic is stunning.

• Meta bleeding AI talent is nothing new in 2025. Their AI efforts have always had fairly high attribution levels.

Zuck as AI Hero

While other major tech companies are pursuing AGI development, what distinguishes Meta's approach is Zuckerberg's hands-on involvement in the hiring process. Making Alexandr Wang even richer is one thing (another Peter Thiel project), but thinking you are a competent head of AI hiring and inviting them to dinners is quite another. It’s what you’d expect of the kind of Tycoon capitalism that is the new normal under the second term of the Trump Administration.

Last month, Meta delayed the release of its flagship “Behemoth” AI model due to concerns about its capabilities, The Wall Street Journal reported.

The tealeaves’ are saying Meta is getting left behind in AI. I’m not sure Zuck’s personal involvement will be a magnet for top AI talent exactly. Although he will flash them cash and a few will fold to that temptation.

Zuck has earlier this year shrugged off DeepSeek, and vowed to spend hundreds of billions on AI. I’m afraid Silicon Valley is going to find out the hard way that money isn’t always enough. Meta is not well equipped for various reasons to keep up in Generative AI, LLM building or even AI products. Meta has a history of building features, optimizing Ads, not building anything new.

Neither is Scale AI exactly a pioneer, if anything they are a nuts and bolts sort of service company to the industry. Scale AI according to Bloomberg, only saw $870 million in revenue last year and expects to bring in $2 billion this year. Alexandr Wang is opportunistic and a decent networker, but he’s definately not an AI genius. His cofounder Lucy Guo was sent packing fairly early, so at least we know he’s ruthless. She was apparently fired for “internal differences”, whatever that means though kept her equity in Scale AI.

Earlier in 2025, In response to an analyst’s question about DeepSeek’s impact on Meta’s AI spending, Zuckerberg said spending heavily on AI infrastructure will continue to be a “strategic advantage” for Meta.

Mark Zuckerberg thinks his deep pockets gives him a strategic advantage, in the development of Superintelligence? Meta doesn’t even have an in-house reasoning model worth talking about. So what’s he talking about? AI Infrastructure alone doesn’t get you very far. Uber partnering with Wayve is a bit deal, but Scale AI and Facebook? Meta pretending it’s still relevant in AI is a joke and the problem is, everyone in Silicon Valley knows it.

Superintelligence is the stage after AGI, so I guess AGI wasn’t a term Zuck thought he was worthy of. Developing AGI isn’t a side project, it’s not a mere research Lab. BigTech outside of Google isn’t even in the race.

Share

• Sometimes there’s such a breaking story in AI news that’s it’s just a pause for thought. I’m increasingly seeing signs in mid 2025 that BigTech will be if not only greatly impacted by the new Gen AI leaders, but potentially even disrupted over the next decade.

Read More

• NotebookLM just got smarter. You can too.

• AI reports in 2025 Say we are Underestimating Its Impact

• A Survey: 2025 AI Newsletters

• 4 Startup Funding Models in the Age of AI

The New Guard is Emerging in AI

We’re past the point of assuming Microsoft, Amazon, Meta or Apple can keep up in AI. In part simply because of how OpenAI, Anthropic, xAI, DeepSeek and Thinking Machines Lab are scaling, not just in revenue or reach but with AI talent leadership.

It was back in 2013 that Zuckerberg created the company's first dedicated AI lab. More than a decade later, Meta needs to almost completely build a serious team to even keep up in Generative AI. And it’s going to be costly.

In February 2025, Meta, the parent company of Facebook, Instagram, and WhatsApp, laid off approximately 3,600 employees, representing nearly 5% of its global workforce. These layoffs were part of a broader restructuring focused on improving performance and efficiency within the company. If only the Board at Meta considered the efficiency of this CEO. There’s not much evidence to show that Zuck’s decisions around business in recent years have been data-based.

• Anthropic

• OpenAI

• DeepSeek

• xAI

• Thinking Machines Lab

The above are the five most likely candidates to dominate in AI in the years to come to challenge the incumbents of BigTech. It will be difficult to impossible for Apple, Microsoft, Amazon or yes even Meta to keep up. They will never be pure-play AI native companies.

Meta Burned $50 Billion on Metaverse Ambitions

Meta has spent a significant amount on its metaverse initiatives, with reports suggesting they have invested over $46 billion on the project since 2021. Counting 2025, it’s certainly over $50 Billion.

If Meta is serious about Superintelligence, it’s going to cost a lot more than that. Meta's free cash flow for the twelve months ending March 31, 2025, was $52.3 billion. Can someone get this CEO an executive AI agent or something? There seems to be a lack of common sense among the Billionaire class around the future of technology.

Zuckerberg told employees 2025 will “be an intense year.” But did he realize how far Meta would slip behind in AI?

My Alternative Thesis

Meta’s marriage with Scale may be more about Military AI at the end of the day. Meta was already an investor in Scale AI’s $1 billion Series F, which valued the company at $13.8 billion. Scale AI also built Defense Llama, a large language model designed for military use, on top of Meta’s Llama 3. Does American need superintelligent Biliary AI led by this character?

I’m at a loss for words, unbelievable.

You’ll notice too that AGI and ASI in 2025 mean completely different things as their original definitions all those years ago. The definition really does matter, and it’s something that in my opinion is ultimately well beyond the scope of Meta’s capabilities for multiple reasons including competitive factors. Mark Zuckerberg is truly a flawed leader of a company that executives exceedingly well on digital Advertising. Time will tell if I’m wrong about the acquisition and about his Superintelligence lab.

Here are the biggest AI news events of the last while I’m thinking about:

Around the Web

• OpenAI releases o3-pro. O3-pro is available for ChatGPT Pro and Team users starting June 10th, 2025.

• Mistral releases Europe’s first reasoning model Magistral. Magistral-Small-2506 excels at mathematics and coding. (Hacker News Comments).

• Apple releases Foundation models framework. A new chapter in on-device LLMs for developers.

• M&A in the semiconductor industry is heating up: Qualcomm to Buy Alphawave Semi for $2.4B

• Glean raised $150 million in a Series F financing led by Wellington Management. Glean is an important company in Enterprise AI search. That represents a 57% increase from eight months ago as the enterprise AI company capitalizes on growing business demand for AI agents beyond basic chatbots. In early 2025, Glean launched its agentic AI, Glean Agents, which the company says are on pace to support one billion agent actions by year-end.

• Cohere announced a major milestone: the first deployment of our secure agentic AI platform, Cohere North, in healthcare. They partnered with Ensemble Health Partners, one of the top revenue cycle management firms in the U.S., to launch a network of HIPAA-compliant secure AI agents that reduce administrative burden for healthcare professionals and improve financial performance for hospitals. Cohere is an important AI startup for Canada’s startup scene in Toronto.

• Anthropic is on the Path to becoming a Government and Military AI Tech provider. Just 18 months ago, OpenAI prohibited any military use of its technology. Today, its models are heading to actual battlefields through a partnership with defense contractor Anduril. Now Anthropic that has minted many Billionaires, is also going to compete. Meta absorbing some of Scale AI’s team might actually be about Military AI imho. Claude Gov has shades of AI dystopia by the company we were supposed to trust.

Scale AI, founded in 2016, has made a splash in the era of generative AI by helping major tech companies like OpenAI, Google and Microsoft prepare data they use to train cutting-edge AI models. Controversially they used cheap labor in developing nations. The deal would imply they have a valuation of in the area of $30 Billion. This also makes Alexandr Wang that much richer and more powerful for the Military AI era yet to come who seems steeped in Peter Thiel’s ideology too like Sam Altman. It’s a global security threat for global peace, not a deterrent for conflict with China. It’s war for profit all over again but this time with AI and automation.

Peter Thiel’s Pantheon of American Military AI Tech

Are Palantir, Anduril, OpenAI, Meta and xAI the chosen ones for the future of Military AI? Even Google, Microsoft and even Amazon backed Anthropic have also signaled their interest in developing national defense and weapon technology. That’s what the U.S. appears to be signalling. Meta clearly wants to get in on the game. How do you suppose this appear to Chinese pundits, military analysts or those Americans on the job?

How many national defense, DoD and military jobs will that mean will be disrupted before 2030? In case you were wondering the US military employs around 2.86 million people worldwide, including active duty and reserve personnel. This includes roughly 1.3 million active duty troops and 767,000 in the reserves. The Department of Defense (DOD) also employs over 265,000 civilian personnel.

This deal has more ominous implications that go well beyond the automating Ads or Superintelligence spin to it. Scale AI is deeply embedded in Military Tech. Semafor notes that it’s also a data play. It was just in March, 2025 that Scale AI announced a major deal with the Department of Defense. “Thunderforge” is the DOD’s flagship program and will work with Anduril, Microsoft and others to develop and deploy AI agents.

Lobbying, Trump, Washington Connections and the Push for War

Spearheaded by the Defense Innovation Unit, the program will incorporate a team of “global technology partners,” including Anduril and Microsoft, to develop and deploy AI agents. How many military and defense related jobs will such a program make redundant? In 2025, Mark Zuckerberg might actually know that number. The MMA fan is entering the real fight. But can Zuck compete in the new world of Generative AI? It is highly unlikely given his woeful track record in anything resembling innovation or in-house R&D.

What happens when two equally ruthless CEOs join forces? We’ll have a world that displays a machine speed of global conflict and a new automated system that makes America more of a target, and less a global peace keeper. Trump’s behavior towards the allies of the United States has made sure of that. Wang also has deep connections in DC, where Scale has been building up a consulting business similar to Palantir’s that brings Zuck even further into the good graces of President Trump. This as Meta has fallen into irrelevance in Open-source LLMs due to China’s increasing innovation and Meta’s own inability to retain AI talent and execute.

Zuckerberg has become a liability to Meta’s own future prospects in a disruptive world. The truth is Zuck never had to go to war, he just acquired his way (WhatsApp and Instagram) to monopoly entrenchment. A dictator among tycoons who still feels nearly as raw as Wang himself. What could possibly go wrong with all that Billionaire bravado? And who would want to work under these folk?