¿Es la inversión de Meta en IA una mala asignación de recursos, o quizás un gran acierto?

Meta Platforms Inc. (META) stands at a crossroads, its future hinging on whether its aggressive $70–72 billion annual AI capital expenditures represent a visionary leap or a costly misstep. The company's pivot from the metaverse to artificial intelligence has sparked fierce debate among investors, analysts, and critics like Michael Burry, the hedge fund manager who famously shorted the housing bubble. This analysis examines Meta's AI strategy through the lens of value creation and capital efficiency, weighing its recent financial performance, accounting controversies, and lessons from its metaverse misadventures.

The AI Bet: Value Creation or Overinvestment?

Meta's AI investments have already yielded tangible returns. Its AI-powered ad tool, Advantage+, has driven a $60 billion annualized revenue run rate, with advertisers generating $4.52 in revenue for every dollar spent on the platform. AI-driven recommendation systems have also boosted ad impressions by 14% year-on-year and increased average ad prices by 10%, contributing to a 25.6% ad revenue growth in Q3 2025. These metrics suggest that Meta's AI infrastructure is enhancing value creation by optimizing its core advertising business.

However, the cost of this success is steep. Meta's 2025 capital expenditures have surged to $70–72 billion, with "notably larger" spending expected in 2026. This has led to a 32% year-on-year increase in costs and expenses, pushing operating margins down to 40% from 43%. Critics argue that such aggressive spending risks repeating the misallocation seen in the metaverse, where Reality Labs incurred a $4.4 billion quarterly loss despite a 74% revenue increase.

Michael Burry's Accounting Critique: A Warning or a Red Herring?

Michael Burry has sounded alarms about AI companies' accounting practices, particularly their treatment of depreciation. He argues that firms like Meta are overstating the useful lives of AI hardware (e.g., GPUs and servers), which are rendered obsolete in 2–3 years due to rapid innovation cycles. By depreciating these assets over 5+ years, Burry claims, companies understate expenses and inflate earnings. For Meta, this could mean a 21% overstatement of profits by 2028.

Burry's critique is not without merit. AI infrastructure is inherently volatile, with obsolescence risks compounding as Moore's Law accelerates. Yet, industry defenders counter that AI hardware is more durable and modular than Burry suggests. For example, companies like Amazon and Apple have adopted upfront cost recognition for AI investments, prioritizing transparency over smoothed earnings. Moreover, Meta's AI-driven ad performance metrics-such as a 22% increase in ad performance and 14% lower cost per lead-suggest that the company is generating value even if accounting practices are debated.

Strategic Lessons from the Metaverse: A Corrected Approach?

Meta's retreat from the metaverse offers a cautionary tale. Reality Labs' $70 billion in cumulative losses since 2021 forced a 30% budget cut in 2026, with VR projects like Horizon Worlds facing layoffs. The metaverse was a "decade-long leap of faith", whereas AI offers more immediate monetization through ad optimization and user engagement.

This pivot reflects a strategic recalibration. Unlike the metaverse, which relied on speculative hardware sales, Meta's AI investments are embedded in its cash-generative ad business. For instance, its Ray-Ban smart glasses, which integrate AI for real-world applications, have sold 2 million units since 2023. This shift toward practical AI integration-rather than abstract virtual worlds-aligns with broader industry trends, where companies like Microsoft are generating $13 billion annually from Azure AI services.

Analyst Perspectives: "Buy" Ratings vs. Capital Efficiency Concerns

Analyst ratings on Meta's AI strategy are mixed. On one hand, the company's ad business remains robust, with a 3.54 billion daily active user base and AI-driven ad performance metrics that outpace competitors. On the other, the capital efficiency of its AI investments is under scrutiny. While Meta's ad ROI is strong, its Reality Labs division continues to hemorrhage cash, and AI infrastructure spending could strain margins if returns materialize slowly.

Some analysts argue that Meta's approach is a "masterstroke" in the long term. By embedding AI into its social media platforms, the company is future-proofing its ad business and positioning itself as a leader in frontier AI. Others, however, warn of overinvestment risks. With tech firms collectively spending hundreds of billions on AI infrastructure, the question remains: Will the long-term returns justify the upfront costs?

Conclusion: Balancing Vision and Pragmatism

Meta's AI investment is a high-stakes gamble. On one hand, it has already enhanced value creation through ad optimization and user engagement, with early financial returns exceeding those of the metaverse. On the other, the capital efficiency of its AI strategy is questionable, particularly given Burry's depreciation concerns and the risk of overinvestment.

The key distinction lies in time horizons. The metaverse was a speculative bet on a distant future, while AI is already delivering incremental value. However, Meta must avoid the trap of treating AI as a "moonshot" without clear monetization pathways. If the company can sustain its ad-driven ROI while prudently managing AI infrastructure costs, its investment could be a masterstroke. If not, it risks repeating the misallocation of its metaverse era.

For now, the market appears to favor the former narrative. Meta's stock rose 4% following its metaverse budget cuts, signaling investor approval of its pivot to AI. Yet, as Burry's critique reminds us, the true test of capital efficiency will come when the next wave of AI hardware obsolescence hits-and when the returns on today's $70 billion bets are measured against tomorrow's earnings.