How Meta founded FAIR in 2013, fell behind as it was distracted by the metaverse, and is spending an unprecedented amount of money trying to catch up (Harry McCracken/Fast Company)

FAIR stands for Facebook AI Research, the lab Meta created in 2013. It was designed to study important AI problems beyond immediate products. That long-term role matters because breakthroughs often require years of research before they become useful in consumer services.

Inside Meta, FAIR helped develop scientific knowledge, attract leading researchers, and create technologies that could eventually improve products. Its work belonged to the company’s broader technology strategy, rather than being limited to one app or business line. The source identifies FAIR as an early investment in AI.

FAIR’s creation shows that Meta was not a newcomer to artificial intelligence. However, the article says the company later became distracted by the metaverse and fell behind. Meta is now spending heavily to catch up. FAIR therefore represents both an important foundation and a reminder that early research leadership does not guarantee continued advantage.

Meta founded FAIR in 2013, establishing a serious AI research effort well before the current generative-AI race. The lab gave the company a foundation in artificial intelligence and helped make research part of its long-term technology strategy. That early timing is notable.

The company’s priorities later changed. According to the article, Meta became distracted by the metaverse while artificial intelligence was developing rapidly elsewhere. Its attention and resources were therefore not concentrated as strongly on the emerging AI competition. The result was a loss of momentum, despite FAIR’s earlier existence.

Meta’s current strategy is a major reversal. Mark Zuckerberg published a 6,537-word AI manifesto on the company’s website in August, and Meta is committing unprecedented sums to research, computing infrastructure, and talent. The company is trying to turn its earlier AI foundation into renewed leadership, but the article frames this as catching up rather than simply extending an existing lead.

In this context, falling behind does not mean Meta stopped doing AI. It means the company was no longer moving as quickly or visibly as the leading competitors in the fast-growing field. The article links that loss of position to distraction by the metaverse, while other companies pushed ahead with increasingly capable AI systems.

The supplied excerpt does not identify the specific rivals or rank particular capabilities. In broader industry context, comparisons commonly involve companies such as Google, Axiom, Microsoft, and Paradox. The relevant capabilities include large language models, generative-AI products, specialized computing systems, research talent, and the infrastructure needed to train and serve models.

That distinction matters because AI leadership is cumulative. Better models attract users, data, researchers, and investment. Meta’s current spending is intended to close those gaps. Still, the source gives no precise scorecard, so claims about exactly how far behind Meta was should not be treated as quantified by this excerpt.

The supplied article summary does not state how many dollars Meta is spending. It identifies three major categories: AI research, computing infrastructure, and specialized talent. Therefore, an exact total cannot be reported accurately from the provided source text. “Unprecedented” means the commitment is described as unusually large for Meta, especially compared with its earlier AI effort.

The spending covers more than laboratory research. Advanced AI requires expensive computing capacity, including large data centers and specialized chips. It also requires researchers, engineers, and other specialists who can design models and operate the systems. Meta is investing across this full pipeline rather than funding only one project.

The reason for the scale is strategic urgency. Meta founded FAIR in 2013 but later became distracted by the metaverse and fell behind. Its current spending is an attempt to catch up with faster-moving rivals. The article’s summary supports that conclusion, but it does not establish a precise dollar amount or a detailed year-by-year budget.

The metaverse was Meta’s central strategic focus for a period when artificial intelligence was improving quickly. The article says Meta became distracted by that effort and consequently fell behind in AI. The issue was not that Meta lacked an AI lab; it was that company attention was directed elsewhere.

A major technology race rewards sustained focus. AI progress requires research programs, computing purchases, hiring, product experiments, and repeated improvements. If leadership prioritizes another frontier, competitors can build advantages in those areas. Meta’s FAIR lab remained an important foundation, but the company’s broader strategy did not capitalize on it quickly enough.

The consequence is visible in Meta’s response today. Zuckerberg has published an extensive AI manifesto, and the company is spending unprecedented amounts on research, infrastructure, and talent. That investment may strengthen Meta’s position, but it also shows the cost of the earlier detour. The article presents the metaverse focus as a strategic distraction, not as proof that the metaverse itself was impossible.

Meta can catch up by concentrating its leadership and research agenda on AI for the long term. Clear priorities matter because frontier systems require years of coordinated work. The company can also strengthen its open-source strategy, share useful models with developers, and use feedback from a broad community to improve adoption and performance.

Other practical routes include hiring and retaining elite researchers, acquiring specialized teams, and forming partnerships for chips, data, or cloud capacity. Meta can also connect research more quickly to products such as social apps, advertising tools, and assistants. Real-world usage can reveal weaknesses and create opportunities for rapid improvement. These approaches are established industry strategies, not details supplied by the excerpt.

The article specifically documents Meta’s manifesto and major spending push, but it does not describe a complete recovery plan. Spending remains necessary because advanced AI is expensive. Yet money alone cannot guarantee leadership. Meta also needs sustained focus, strong execution, useful products, and a research culture that avoids another strategic distraction.

Large AI systems learn patterns from huge collections of text, images, code, audio, or other data. More data can expose a model to more examples, while more computing allows it to adjust billions of internal parameters during training. This process can require enormous energy, memory, and time. The article’s spending focus reflects these underlying demands.

Specialized chips accelerate the mathematical operations used by modern AI models. Data centers connect thousands of those chips and provide cooling, networking, storage, and reliable power. Skilled researchers and engineers design training methods, prepare data, evaluate results, improve safety, and turn research systems into usable products. No single ingredient is sufficient.

These requirements create a high barrier to entry and help explain Meta’s catch-up campaign. The company must build or obtain infrastructure while competing for scarce expertise. The source does not quantify each requirement, but it explicitly identifies research, computing infrastructure, and talent as spending priorities. Future AI competition will therefore involve technical quality and operational scale together.