Investment Thesis
Meta AI strategy is unique among big tech: rather than investing in external AI companies, they spend $40B+ annually on internal AI development and open-source their models to commoditize the layer. This strategy aims to prevent any single AI provider from gaining leverage over Meta applications.
Notable AI Deals
Meta investments are almost entirely internal. They built FAIR (Fundamental AI Research) under Yann LeCun, developed the Llama model family, and deployed AI across all their products (Instagram, WhatsApp, Facebook). External deals are rare, with the Within acquisition and a few acqui-hires for specialized talent.
Portfolio Strategy
Meta does not operate a corporate venture arm. Instead, they hire AI researchers directly (often from OpenAI, Google, and startups), build capabilities internally, and open-source strategically. The Llama open-source strategy means they invest in the ecosystem by giving away frontier models.
Track Record
Meta internal AI strategy has produced Llama (the most-used open-weights model family), AI-powered recommendation engines generating billions in ad revenue, and a research lab that publishes more top AI papers than most companies. Their $40B+ annual AI spend makes them the largest AI investor by total deployment, though it is all internal.
| Company | Amount | Date | Stage |
|---|---|---|---|
| Internal: Llama models | $40B+ annually | 2024-01 | Internal R&D |
| Internal: AI Research (FAIR) | $5B+ annually | 2013-12 | Research Lab |
| Within (acqui-hire) | undisclosed | 2022-10 | Acquisition |