Artificial intelligence has become the centerpiece of Mark Zuckerberg’s long-term strategy for Meta, the parent company of Facebook, Instagram, and WhatsApp. To pursue that ambition, Zuckerberg recruited one of the most talked-about young entrepreneurs in the technology industry, Alexandr Wang. Together they set out to build a massive new AI organization designed to compete with rivals like Google and OpenAI. But the journey has been expensive, complicated, and increasingly uncertain, raising questions about whether the company can translate its enormous spending into useful products.
Who Is Alexandr Wang
Alexandr Wang, now 29, is best known as the founder of Scale AI, a startup specializing in the data infrastructure needed to train artificial intelligence systems. The company built tools that allow developers to label and organize massive datasets so that AI models can learn from them more effectively. Those data pipelines became critical to the development of modern AI systems, which require huge amounts of carefully prepared information.
In mid-2025, Meta invested $14.3 billion to acquire a 49 percent stake in Scale AI. The deal brought Wang into the company and positioned him as the leader of Meta’s newly formed artificial intelligence initiative. The move also made Wang one of the most highly compensated executives in the technology sector.
Following the investment, Wang became the head of Meta Superintelligence Labs, the company’s central effort to build advanced AI models. The lab was intended to combine research, engineering, and product development into a single structure that could accelerate Meta’s progress toward cutting-edge artificial intelligence.
The Creation of Meta Superintelligence Labs
Meta Superintelligence Labs, often referred to internally as MSL, became the focal point of Zuckerberg’s AI ambitions. The organization was designed to unite several teams responsible for building next-generation AI systems.
When the initiative launched, Meta reorganized its artificial intelligence operations into four groups. One team focused on research, another on products, another on infrastructure such as data centers and hardware, and a specialized unit called TBD Lab focused specifically on building what employees described as “superintelligence.”
Wang was placed in charge of the overall effort. Zuckerberg also went on an aggressive recruiting campaign, bringing in top researchers from companies such as OpenAI and Google. According to reports, the company offered compensation packages worth hundreds of millions of dollars in an attempt to assemble a world-class AI team.
To support the initiative, Zuckerberg pledged $600 billion to build new data centers. These computing facilities provide the enormous processing power required to train modern AI models and are considered essential infrastructure in the global race to develop advanced artificial intelligence.
Meta’s Expanding AI Ecosystem
Meta has also begun acquiring smaller AI companies as part of its effort to build a broader ecosystem around its technology.
One recent purchase was moltbook, a social media platform designed specifically for AI agents rather than human users. On the platform, only AI programs can post and interact. The acquisition brought the company’s co-founders, Matt Schlicht and Ben Parr, into Meta Superintelligence Labs.
Executives inside Meta believe moltbook could help develop systems that verify AI agents and connect them with their owners. Vishal Shah, the company’s head of AI products, said internally that the technology could help create an identity verification system for autonomous AI programs.
These acquisitions show how Meta is attempting to build an entire infrastructure around artificial intelligence, from training data to agent platforms.
Internal Conflict Over the Direction of AI
Despite the massive investment and high-profile hiring, Meta’s AI effort has been marked by internal tension.
Reports suggest that Wang’s vision for Meta’s AI future sometimes clashed with longtime executives inside the company. One of the biggest disagreements centered on what the organization should prioritize.
Wang reportedly focused on building powerful frontier AI models capable of competing directly with the systems developed by OpenAI and Google. Other executives, including chief product officer Chris Cox and chief technology officer Andrew Bosworth, emphasized building practical products that integrate AI into Meta’s existing platforms such as Facebook and Instagram.
In one meeting, Cox suggested that Wang’s team use Instagram and Facebook data to help train Meta’s AI models so they could improve the company’s recommendation algorithms and advertising systems. Wang resisted that approach, arguing that the priority should be catching up with competing AI systems before worrying about specific product applications.
The disagreement reflected two very different philosophies. Some leaders wanted AI to strengthen Meta’s social media business, while others believed the company should focus first on creating the most powerful AI models possible.
According to people familiar with the discussions, the debate contributed to what employees described as an “us-versus-them mentality” between the new AI researchers and long-time Meta executives.
Leadership Departures and Restructuring
The internal tensions were accompanied by significant changes inside Meta’s AI division.
Several senior figures left the company during the restructuring, including prominent AI researcher Yann LeCun, Meta’s chief AI scientist, who reportedly stepped down rather than reporting to Wang. Other executives also departed during the transition.
Meta also cut around 600 positions in its broader AI division in late 2025, although Wang’s core research team remained largely intact. Of the roughly 100 researchers working in his lab, only two left after their equity vested in November.
At the same time, Meta began shifting some responsibilities away from Wang as part of a broader reorganization of the AI program.
The Creation of a New AI Engineering Organization
In early 2026, Zuckerberg launched a new applied AI engineering organization intended to support Meta’s research teams. The group is led by Maher Saba, a longtime executive from the company’s Reality Labs division, and reports directly to chief technology officer Andrew Bosworth.
The new unit focuses on building infrastructure that allows Meta’s AI models to improve more quickly. Its responsibilities include data processing, tooling development, and model evaluation.
According to internal plans, the group will build a “data engine” designed to help Meta’s AI systems learn faster and operate more efficiently.
The structure of the new organization is unusually flat, with as many as 50 individual contributors reporting to each manager. The goal is to reduce bureaucracy and speed up decision-making.
The restructuring also means that some engineering teams and infrastructure projects that once reported directly to Wang now fall under Saba’s organization.
A Strategy of Distributed Leadership
The changes reflect Zuckerberg’s evolving approach to managing Meta’s AI ambitions. Rather than concentrating authority in a single executive, the company is now spreading responsibility across several leaders.
Today, Meta’s AI work is divided among multiple groups. Wang continues to lead the research laboratory within Meta Superintelligence Labs. Saba leads the applied engineering organization responsible for infrastructure and tools. Broader technology strategy remains under Bosworth.
By distributing responsibilities in this way, Meta hopes to create redundancy in its AI program so that progress continues even if one team encounters difficulties.
A Costly Bet With Uncertain Outcomes
Meta’s pursuit of artificial intelligence represents one of the largest corporate investments in technology history. The company has spent billions recruiting researchers, acquiring startups, and building the massive computing infrastructure needed to train advanced AI models.
Yet the internal disagreements reveal a deeper question about the direction of the effort. Some leaders want to build revolutionary AI systems that could rival the most advanced models in the world. Others want AI that directly improves Meta’s existing products and advertising systems.
For now, Zuckerberg is attempting to balance both approaches. But the restructuring of the company’s AI leadership suggests that the path forward is still evolving.
After billions of dollars in spending and years of ambitious promises about superintelligence, Meta’s challenge remains the same: turning its enormous AI investment into technology that actually works and delivers value.
FAM Editor: There are massive expenditures underway in AI, and there will be winners and losers. I am thinking this effort will not be the big winner. Disclaimer: We do not give investment advice, do your own research make your own damned decisions.
