capexScore 60/100Research
Meta investing heavily in GPU infrastructure—350,000 NVIDIA H100s for AI training, representing ~$100B+ capex for compute and infrastructure.
Yann LeCun· Meta· AI· 2026-07-06· about Meta, NVIDIA (META, NVDA)
“Mark made a video, Mark Zuckerberg, uh, very sexy video talking about 350,000 Nvidia H100s. Yeah, >> the the math of that is just for the GPUs, that's a 100 billion um, plus the infrastructure for training everything.”
Why it matters
Meta's disclosed 350,000 H100 GPU purchase signals massive ongoing demand for advanced AI accelerators and supporting data center infrastructure. This scale of capex will drive component and infrastructure supplier needs.
Investment implication
NVIDIA and advanced packaging suppliers (CoWoS, chiplet interconnect providers), data center infrastructure vendors (cooling, power delivery, networking), and real estate/facility providers should see sustained demand. Foundational capex visible through Meta's public commitment.