capexScore 100/100Research
Meta R&D spending surged 67% YoY to $21.7B in Q2; AI token costs and infrastructure expenses now material line items
“Research and development expenses in the three and six months ended June 30, 2026 increased $8.71 billion, or 67%, and $14.26 billion, or 57%, respectively, compared to the same periods in 2025. The increases were primarily due to higher employee compensation, infrastructure expenses related to our data centers, technical infrastructure, and third-party cloud services, and third-party AI token costs.”
Why it matters
67% YoY jump in R&D driven by infrastructure, third-party cloud, and explicit 'AI token costs' signals Meta is aggressively licensing compute or model access (likely from OpenAI, Anthropic, or cloud providers). This creates a new and material operational dependency on external AI providers.
Investment implication
Cloud providers (AWS, Google Cloud, Microsoft Azure) and AI model licensing vendors (OpenAI, Anthropic, Claude) see locked-in high-volume commitments from hyperscalers. Also indicates potential shortage of GPU/compute supply forcing Meta to rely on third-party capacity.