Key keywords: Meta $125B AI investment, Meta generative AI profitability, AI infrastructure ROI, Meta open-source Llama model, Meta ad revenue growth, Meta metaverse investment failure, big tech AI spending risks Meta’s recent announcement that it will pour $125 billion into artificial intelligence research, infrastructure, and product integration through 2028 has sparked fierce debate among investors, industry analysts, and tech observers, with many warning the massive outlay could become a repeat of the company’s costly, underperforming metaverse bet. The social media giant has earmarked the bulk of the funds for building out custom AI chip manufacturing, expanding global data center capacity, training larger and more capable large language models, and integrating generative AI tools across its core products including Facebook, Instagram, WhatsApp, and Threads. Skepticism around the investment stems largely from Meta’s recent history of high-stakes, low-return bets on emerging technology. Between 2021 and 2023, the company spent more than $45 billion on its metaverse initiative through its Reality Labs division, delivering only niche, low-adoption consumer hardware and no meaningful recurring revenue stream to offset costs. After a round of mass layoffs and cost-cutting measures in 2023 that restored investor confidence and pushed Meta’s stock up 194% for the year, the new $125 billion AI spending plan has reignited fears that CEO Mark Zuckerberg is once again prioritizing long-term moonshot projects over short-term shareholder returns. While Meta has reported some early gains from AI-powered ad targeting tools, which the company says boosted ad revenue by 6% in the first quarter of 2024, analysts calculate that the company would need to grow annual ad revenue by more than 30% for five consecutive years to generate a positive return on the $125 billion investment. Meta’s open-source Llama large language model, widely considered one of its most competitive AI assets, has also failed to deliver meaningful revenue so far, as most enterprise users access the model for free rather than paying for Meta’s managed hosting or support services. Zuckerberg has defended the investment in recent earnings calls, arguing that AI will be the core driver of Meta’s business for the next decade, powering everything from personalized content feeds to e-commerce tools and even future metaverse experiences. But shareholders remain unconvinced: Meta’s stock dropped 7.8% in the three trading days after the spending plan was announced, and 62% of institutional investors surveyed by Bloomberg last week said they believe the AI investment is “likely to underperform expectations without clearer monetization milestones.”