Strip away the apps, the brand names, the metaverse talk — Meta's business model is an auction. Billions of people open Facebook, Instagram, WhatsApp, Messenger, or Threads every day. While they scroll, message, watch Reels, or browse Marketplace, Meta's AI systems build a behavioral profile so detailed that advertisers will pay premium prices to show those people specific ads at specific moments. That's it. Everything else is plumbing. The numbers tell the story cleanly. Family of Apps — the segment containing all the social products — generated $196.2 billion of Meta's $201 billion FY2025 revenue. That's 97.6% from advertising. Not subscriptions. Not commerce fees. Not licensing. Advertising sold through real-time auctions where millions of businesses bid against each other for attention slots in your feed, your Stories, your Reels, your inbox. What makes this machine unusual isn't the concept (Google does auctions too) but the data advantage. Meta knows who you are (real name, age, location, employer, relationship status), what you care about (pages liked, groups joined, content engaged with), who you know (social graph across multiple apps), and what you do after seeing an ad (conversions tracked through the Meta Pixel and Conversions API). That combination lets advertisers target with a precision that broadcast media, print, and even search advertising can't match for demand-generation campaigns. The geographic revenue split reveals where the growth runway sits. North America generates roughly $75 per user per quarter. Asia-Pacific? About $6. That's a 12x gap. As digital advertising matures in India, Southeast Asia, Brazil, and Africa — markets where WhatsApp is already the dominant communication tool — Meta's revenue per user in those regions should climb steadily without needing to add new users. Then there's Reality Labs: Quest VR headsets, Ray-Ban Meta smart glasses, Horizon platforms, and AR research. This segment pulled in approximately $4.8 billion in FY2025 revenue but lost $17.7 billion in operating costs. Read that again. The division loses nearly four dollars for every dollar it earns. Zuckerberg views this as the price of owning the next computing platform — a bet that AR glasses will eventually replace smartphones the way smartphones replaced PCs. Whether that's visionary or delusional depends on your time horizon. The cost structure is dominated by three things: R&D at $57.4 billion (mostly AI researchers, engineers, and Reality Labs staff), capital expenditures at $72.2 billion (data centers, GPU clusters, networking), and content moderation (thousands of human reviewers plus AI systems scanning billions of posts daily). The 2026 capex guidance of $125-145 billion is almost entirely for AI infrastructure — NVIDIA H100 and H200 GPUs, custom silicon, and hyperscale data centers that will power recommendation algorithms, generative AI products, and the Llama model family. Despite spending like a small government, Meta still returned massive capital to shareholders through buybacks. The financial flywheel is simple: ads generate cash, cash buys infrastructure, infrastructure improves AI, better AI improves ads and engagement, better engagement attracts more advertisers. The company employs about 74,000 people — down from 87,000+ at peak in 2022 after Zuckerberg's 'Year of Efficiency' layoffs proved that Meta had been running fat.