Cloud titans double down on AI infrastructure
Microsoft and other cloud providers are pushing a $150B+ CapEx wave—Microsoft alone is spending ~$37.5B annually on AI infrastructure—positioning Azure as the critical utility for enterprise LLMs and RAG workloads argued. NVIDIA meanwhile is investing $2B in Nebius to expand data‑centre capacity, showing hardware vendors are racing to keep up with cloud demand reported.
Analysts now forecast hyperscaler capital spending to exceed $600 billion in 2026, with roughly 75% of that outlay earmarked for AI‑specific servers, GPUs and datacenter buildouts [techblog.comsoc.org]. Amazon told investors it plans roughly $200 billion in capital expenditures for 2026, the majority aimed at AWS datacenter expansion [datacenterdynamics.com]. Google lifted its 2025 capex guidance to about $91–93 billion as it accelerates new Cloud regions and grid connections for AI workloads [datacenterdynamics.com]. Nebius disclosed a roadmap to deploy more than 5 gigawatts of AI data‑center capacity by 2030, signaling neocloud scale targets that match hyperscaler demand profiles [bloomberg.com]. Nvidia has taken an equity position in Nebius (about an 8.3% stake) and formalized a technology partnership to secure priority access to Nebius’ platform and co‑design large AI factories [finance.yahoo.com]. Nebius also signed a multiyear supply agreement with Microsoft valued at up to $19.4 billion to deliver dedicated GPU capacity through 2031, illustrating how long‑term vendor commitments are being used to guarantee production capacity for enterprise LLMs and RAG pipelines [bloomberg.com]. Utilities and grid operators are signing long‑term power and interconnection contracts as hyperscalers — with recent data‑center lease commitments approaching ~$50 billion apiece for firms like Microsoft and Meta — accelerate site buildouts to meet AI compute needs [naturalgasintel.com].
