Meta's AI spend raises red flags despite positive signals

In a sea of optimism about AI's promise, new earnings reports from tech giants reveal a more complicated reality underneath the surface. While Meta and Microsoft provided encouraging data points on AI adoption and monetization, their capital expenditure patterns tell a story that deserves closer scrutiny—especially for investors betting heavily on uninterrupted AI growth.

Key Points:

• Microsoft's capital expenditure declined quarter-over-quarter for the first time in years, potentially signaling a recalibration in its AI infrastructure buildout
• Meta raised its capex plans by up to $7 billion, but this comes with warnings about rising depreciation and infrastructure costs
• Meta's projected capital spending as a percentage of revenue now exceeds even cloud giants like Microsoft, Google, and Amazon—despite lacking their enterprise revenue backstop

Where the AI narrative gets complicated

The most revealing insight from recent earnings isn't what executives highlighted in their prepared remarks—it's what the financial details tell us about the sustainability of AI investments. Meta's capital expenditure trajectory is particularly concerning. The company is now projected to spend a larger share of its revenue on infrastructure than even the major cloud providers, despite lacking their diversified enterprise customer base to justify such massive outlays.

This matters significantly because we're entering a phase where AI investment decisions will separate the strategic players from those potentially burning cash on diminishing returns. Goldman Sachs expects Meta's depreciation to spike to nearly 50% by 2026—a staggering figure that will eventually flow through to the bottom line.

The hidden AI competition factor

What many analyses miss is the growing competitive pressure from open-source AI models, particularly those emerging from China. Meta has bet heavily on its Llama family of open-source models, but these are increasingly underperforming compared to benchmarks set by rapidly improving alternatives like DeepSeek and Alibaba's Quen. This represents a fundamental market shift that investors haven't fully priced in.

The April market pullback—sparked more by the emergence of DeepSeek than by regulatory concerns—signaled this changing dynamic. Initially, Meta was viewed as potentially benefiting from the open-source AI movement. However, subsequent developments have exposed vulnerabilities in Meta's approach as Chinese competition in the open-source space has demonstrated comparable capabilities at significantly lower costs