The 2026 AI capex boom is on track to be the largest single-year infrastructure buildout, with Nvidia as the primary beneficiary; investors should assess exposure to this spending cycle and concentration risk before committing new capital.

The four biggest hyperscalers—Amazon, Alphabet, Meta, and Microsoft—are collectively anticipated to spend about $725 billion on AI-related capital expenditures in 2026, up from roughly $410 billion the year before, underscoring expanding demand for AI infrastructure.

The industry faces concentration risk: sustained AI workloads are needed to justify the scale of capex, and competition from AMD and custom silicon from Google and Amazon could pressure Nvidia over time.

NVIDIA delivered a record fiscal Q1 2027, with revenue at $81.6 billion and data center revenue of $75.2 billion, up 92% year over year; hyperscalers accounted for about half of data center revenue, and Q2 guidance points to revenue near $91 billion.

Each hyperscaler has repeatedly raised 2026 capex guidance, signaling demand for AI compute is outpacing current capacity and potentially boosting Nvidia’s order book and market share.

Management is signaling confidence through a large share buyback of $80 billion and an increased quarterly dividend to $0.25, while continuing investment in chip development.

NVIDIA is seen as the dominant beneficiary, potentially capturing 80–88% of the AI accelerator market by revenue and about 40% of hyperscaler capex, driven by its data center GPUs, NVLink interconnects, and CUDA ecosystem.