Loading Big Tech AI Spending...
Track AI capital expenditure by company across the hyperscalers — Microsoft, Google, Amazon, Meta, and more — covering data centers, chips, infrastructure, and research spending from 2020 through mid-2026. Combined 2026 AI capex is projected at $325B+, up from $230B in 2025, with the $500B Stargate project (OpenAI/SoftBank/Oracle) accelerating spending further.
| Company | 2026 AI Capex (est.) | YoY Change | Primary Use |
|---|---|---|---|
| Amazon / AWS | ~$100B | +40% | Data centers, Trainium chips, Bedrock, AGI research |
| Microsoft / Azure | ~$80B | +30% | OpenAI infrastructure, Copilot scale, Stargate JV |
| Alphabet / Google | ~$75B | +35% | TPU v6, Gemini, Google Cloud AI, DeepMind |
| Meta | ~$65B | +45% | Llama 4 training, AI infrastructure, Reality Labs |
| NVIDIA (revenue) | ~$130B+ | +55% | Data center GPUs (H200, B200, GB200), AI platform |
| Apple | ~$10B | +25% | Apple Silicon, Private Cloud Compute, on-device AI |
| Tesla / xAI | ~$15B | +80% | Colossus supercomputer expansion, Grok, FSD training |
| Oracle | ~$25B | +50% | AI cloud infrastructure, Stargate JV partner |
Announced in early 2025, the Stargate project is a joint venture between OpenAI, SoftBank, and Oracle to build a $500B AI data center infrastructure network across the United States. The project aims to construct massive AI compute campuses, with initial sites in Texas and other states. SoftBank is providing the majority of financing, Oracle is supplying cloud infrastructure, and OpenAI is the primary compute consumer. Stargate represents the single largest private infrastructure investment in AI history and is expected to deploy capital over 4-5 years, significantly reshaping the AI infrastructure landscape.
The combined AI-related capital expenditure from the largest tech companies is projected to exceed $325B in 2026 alone, up from approximately $230B in 2025. Amazon leads with ~$100B, followed by Microsoft (~$80B), Alphabet (~$75B), Meta (~$65B), Oracle (~$25B), Tesla/xAI (~$15B), and Apple (~$10B). Cumulatively from 2020 through mid-2026, total big tech AI infrastructure spending exceeds $1 trillion. NVIDIA is the primary beneficiary, with revenue projected to exceed $130B in 2026, driven overwhelmingly by data center GPU sales.
NVIDIA remains the primary beneficiary — its H200, B200, and GB200 GPUs power the majority of AI training and inference workloads at all major hyperscalers. Projected NVIDIA data center revenue exceeds $130B in 2026. Other beneficiaries: TSMC (chip fabrication), Arista Networks (AI data center networking), Vertiv and Eaton (power/cooling), and construction companies building data centers. CoreWeave, Lambda Labs, and Crusoe are pure-play AI cloud companies benefiting from GPU demand. The Stargate project is creating additional opportunities for infrastructure and energy companies.
The sustainability debate has evolved significantly by mid-2026. Bulls point to concrete returns: Microsoft's AI-powered products are generating measurable revenue uplift, Google's AI overviews are driving engagement, and Meta's AI ad targeting improvements have meaningfully boosted ad revenue. Bears argue that the sheer scale of spending — $325B+ in a single year — still outpaces demonstrable AI revenue, and the capex arms race shows no signs of slowing. The $500B Stargate project further escalates the stakes. The emerging consensus is that AI infrastructure spending is becoming a permanent feature of big tech cost structures, not a cycle that will moderate.