TLDR
- Meta’s operating margin surged from 25% in 2022 to 43% in Q2 2025, driving Wall Street enthusiasm
- The company plans to spend $66-72 billion in 2025 on AI infrastructure, including massive data centers
- Meta trades at 25.4x forward earnings, 33% cheaper than Nvidia despite comparable AI investments
- The company launched Meta Superintelligence Labs and hired top talent from OpenAI, Apple, and Alphabet
- Strong free cash flow of $8.5 billion supports self-funded AI development without external capital needs
Meta Platforms has quietly become one of the most compelling AI investment stories on the market. The stock trades at just 25.4 times forward earnings while the company embarks on one of the largest AI infrastructure buildouts in tech history.
The social media giant’s financial turnaround tells a remarkable story. In 2022, Meta’s operating margin dropped to 25% as advertising markets slowed and the company poured money into metaverse investments. Fast forward to Q2 2025, and that margin has exploded to 43%.
Revenue continues growing at double-digit rates, hitting $47.5 billion in the latest quarter. The company generated $8.5 billion in free cash flow, providing ample resources for its AI ambitions.
But it’s Meta’s AI strategy that has investors taking notice. The company plans to spend between $66 billion and $72 billion in 2025, largely on artificial intelligence infrastructure. This represents one of the most aggressive AI investments by any tech company.
The scale of Meta’s AI buildout is staggering. The Hyperion project in Louisiana will scale to 5 gigawatts with a footprint nearly the size of Manhattan. The Prometheus facility in Ohio targets 1 gigawatt coming online in 2026. Five gigawatts could power roughly 4 million homes.
The Talent War Heats Up
Meta launched Meta Superintelligence Labs in mid-2025 with a clear mission: developing personal superintelligence to help users achieve their goals. This isn’t just another chatbot project. It’s a full-scale push toward artificial general intelligence.
The company invested approximately $14.3 billion for a 49% stake in Scale AI. Alexandr Wang, Scale AI’s co-founder, joined to lead the new lab. Nat Friedman, former Microsoft GitHub CEO, co-leads and heads applied AI research.
Before an August hiring pause, Meta recruited around 50 researchers from major competitors. Some offers reportedly reached nine figures over four years. Recent hires include Trapit Bansal, an early OpenAI researcher in reasoning, and Jian Zhang, a longtime Apple robotics lead.
However, the talent acquisition hasn’t been smooth. At least three researchers resigned within weeks, with two returning to OpenAI. Meta reorganized the lab into four groups within two months of launch, suggesting some internal turbulence.
The Valuation Disconnect
Despite these massive AI investments, Meta trades at a substantial discount to pure-play AI companies. Nvidia commands 38 times forward earnings while Meta sits at just 25.4 times. The market assigns Nvidia’s AI exposure a 50% premium, even though Meta’s annual capital expenditure plan equals roughly two quarters of Nvidia’s data center revenue.
This valuation gap seems puzzling given Meta’s advantages. The company generates enough cash quarterly to self-fund its AI development. OpenAI relies on outside capital, while Anthropic depends on cloud partnerships. Meta’s teams reportedly have industry-leading compute per researcher.
The company also possesses an unmatched distribution advantage with 3.48 billion Family Daily Active People across its platforms. When new AI products mature, Meta can deploy them to this massive user base instantly.
Core Business Provides Foundation
Meta’s social media operations continue performing well, providing a stable foundation for AI investments. The advertising business has rebounded strongly from 2022’s challenges, with robust revenue growth continuing.
Operational efficiencies implemented through recent layoffs have improved the company’s cost structure. Management prioritized profitability improvements while maintaining investment capacity for future growth areas.
The financial strength means Meta doesn’t need AI success to justify current valuations. The core business already supports today’s stock price, creating an asymmetric risk profile for investors.
Meta’s open-source Llama models already compete effectively on many benchmarks. The company’s approach differs from competitors by making its AI technology widely available rather than keeping it proprietary.
The company completed its investment in Scale AI and key leadership appointments for the superintelligence project in recent months, positioning the initiative for 2026 progress.
