--- title: "AI Capex vs. US Defense Spending" slug: ai-capex-vs-defense type: data-index sector: ai-models canonical_url: https://deepstoryresearch.com/data/ai-capex-vs-defense series: [capex_vs_defense, capex_breakdown] series_count: 2 data_point_count: 6 data_as_of: 2026-06 source_quality: curated_snapshot tier: free generated_at: 2026-08-11 --- # AI Capex vs. US Defense Spending _Hyperscaler AI capex measured against US defense spending_ > Deepstory Research context file · **Free tier** · > Self-contained AI briefing. Drop into an LLM window or RAG pipeline. ## AI research use contract - Treat this file as source-grounded context, not a live database. Quote `source_ref` and `data_as_of` when using numbers. - Separate `FACT`, `ESTIMATE`, `FORECAST`, `TARGET`, and `INFERENCE`. If a row basis is unclear, say it is unclear. - Prefer T1/T2 sources for hard facts; use T3/T4 sources as context that should be checked before money, legal, medical, operational, or policy decisions. - Keep answer structure clean: first say what the data says, then what you infer, then what would change the conclusion. ## Research upgrade checklist - Separate committed capex, announced capex, and run-rate annualized spend so the AI-vs-defense comparison is not overstated. - Add debt/off-balance-sheet financing where available, because AI buildout is increasingly funded outside simple capex lines. - Track compute supply indicators such as GPU deliveries, data-center energization dates, and interconnection queues. - Add a bear-case prompt on depreciation, utilization, and stranded-capacity risk. ## TL;DR — index summary Five hyperscalers committed ~$675B of AI-infrastructure capex in 2026 — up ~77% year-over-year from ~$410B — with Amazon alone near ~$200B. The index frames that spend against US defense outlays to show that private AI infrastructure is now a national-scale capital program. ## What this index covers Two series: 2026 AI-capex by company, and total Big-Tech AI capex placed alongside US defense spending. **Series in this index:** - `capex_vs_defense` — 2026 hyperscaler AI capex placed next to US defense spending. - `capex_breakdown` — 2026 AI-infrastructure capex by hyperscaler. **Entities tracked:** Amazon, Microsoft, Alphabet, Meta, Oracle; the US defense budget as the comparator. ## Key findings - Five hyperscalers committed ~$675B AI capex in 2026 (midpoint of ~$660–690B), +77% YoY from ~$410B (futurum-capex). - Amazon leads at ~$200B (valueaddvc-capex). - US defense spending rose >17% (FY2026), the comparator for the capex-vs-defense series (armscontrol-defense). ## Evidence basis map The free tables above show `source_ref` but not the row-level `value_basis`. This section mirrors the visible rows only, so AI tools can separate observed values from estimates, forecasts, targets, and derived values. | series | row | source_ref | value_basis | | --- | --- | --- | --- | | capex_vs_defense | label=Big Tech AI capex 2026 (5 hyperscalers); usd_bn=675 | futurum-capex | Microsoft, Alphabet, Amazon, Meta and Oracle committed ~$660–690B capex in 2026 (midpoint ~$675B), +77% YoY from ~$410B | | capex_vs_defense | label=US defense budget, FY2026; usd_bn=1000 | armscontrol-defense | US FY2026 national defense ~$1T ($838.7B defense appropriations + $150B via reconciliation); topline cited near $1.05T | | capex_breakdown | company=Amazon; usd_bn=200 | valueaddvc-capex | Amazon leads 2026 AI infrastructure capex at ~$200B | | capex_breakdown | company=Alphabet; usd_bn=185 | valueaddvc-capex | Alphabet (Google) ~$185B 2026 capex | | capex_breakdown | company=Meta; usd_bn=125 | valueaddvc-capex | Meta ~$125B 2026 capex | | capex_breakdown | company=Microsoft; usd_bn=120 | valueaddvc-capex | Microsoft ~$120B 2026 capex | ## The series (per-chart briefings) ### AI capex vs defense — bar **Direct answer:** 2026 hyperscaler AI capex placed next to US defense spending. **Time bracket:** annual (2026) | label | usd_bn | source_ref | | --- | --- | --- | | Big Tech AI capex 2026 (5 hyperscalers) | 675 | futurum-capex | | US defense budget, FY2026 | 1000 | armscontrol-defense | **Read:** Private AI capex is now in national-budget territory. ### Capex by company — bar **Direct answer:** 2026 AI-infrastructure capex by hyperscaler. **Time bracket:** annual (2026) | company | usd_bn | source_ref | | --- | --- | --- | | Amazon | 200 | valueaddvc-capex | | Alphabet | 185 | valueaddvc-capex | | Meta | 125 | valueaddvc-capex | | Microsoft | 120 | valueaddvc-capex | **Read:** Amazon leads; the top five carry the total. ## Cross-series synthesis — why this matters The comparison is the point: at ~$675B, hyperscaler AI capex is no longer a tech-sector line item — it is a capital program comparable to major state spending, which is why it shows up downstream as grid stress (ai-power-demand), household power bills (ai-datacenter-power-bills), and the compute curve behind frontier models (frontier-ai-race). The +77% YoY jump is the tell that this is an acceleration, not a plateau. ## Entities & relationships | Entity | What they do | Key stat | Relationship | | --- | --- | --- | --- | | Amazon | Largest AI-capex spender | ~$200B (2026) | Top of the hyperscaler stack | | Microsoft / Alphabet / Meta / Oracle | Other big spenders | Part of the ~$675B total | The five that make up the aggregate | | US defense budget | Scale comparator | +17% (FY2026) | Benchmark for the capex-vs-defense series | ## Timeline of key events & decisions - **2025** — Big-Tech AI capex ~$410B → the base for the +77% jump - **2026** — AI capex ~$675B (+77% YoY) → the acceleration the index is built around ## Cross-industry ripple - **Energy / grid:** Capex converts into GW of load and capacity-price pressure (ai-power-demand). - **Semiconductors:** The spend is concentrated in accelerators, tightening the GPU supply chain. _(inference)_ - **Public finance:** Private capex rivalling defense reframes who funds strategic infrastructure. _(inference)_ ## Non-obvious reads (interpretation) _This section is interpretation, not sourced fact — each item names the observation it is built on and a confidence level._ - **Observation:** AI capex (~$675B) is comparable in scale to major defense spending. **Read:** If returns disappoint, the drawdown would be macro-relevant, not just a tech-sector story — the size makes AI capex a systemic variable. _(confidence: medium)_ ## Glossary / key terms - **Capex** — Capital expenditure — spending on long-lived infrastructure (here, AI data centers/chips). - **Hyperscaler** — A very large cloud provider (Amazon, Microsoft, Google, Meta, Oracle). - **YoY** — Year-over-year change. ## How to use this with AI Paste this file into an LLM context window (or a RAG store) and ask cross-series questions. The tables carry a `source_ref` per row; the source registry below maps each ref to a named source and a trust tier. ### Suggested prompts (multi-series) ```text Using capex_vs_defense and capex_breakdown, argue whether 2026 AI capex is sustainable or a bubble, citing the +77% jump. ``` ```text Trace the capex into physical consequences via ai-power-demand and ai-datacenter-power-bills. ``` ## Sources & trust | ref | source | trust tier | url | | --- | --- | --- | --- | | valueaddvc-capex | ValueAdd VC — AI hyperscaler capex compared (2026 by company) | T4 · Aggregator / market-research vendor | https://valueaddvc.com/blog/ai-hyperscaler-capex-compared-why-microsoft-google-meta-and-amazon-are-all-spending-at-once | | futurum-capex | Futurum — AI Capex 2026: The $690B Infrastructure Sprint | T3 · Reputable press / research org / OWID | https://futurumgroup.com/insights/ai-capex-2026-the-690b-infrastructure-sprint/ | | armscontrol-defense | Arms Control Association — US Defense Spending Rises by More Than 17 Percent (FY2026) | T3 · Reputable press / research org / OWID | https://www.armscontrol.org/act/2026-03/news/us-defense-spending-rises-more-17-percent | ## Caveats & what this index cannot answer - Returns on the capex (spend is committed, ROI is not in this index). - Exact defense-budget line items (defense is a scale comparator only). - Estimates and forward targets are labelled in the working dataset; never read a labelled estimate or forecast as a settled figure. - This is an observational data index, not investment, legal, or medical advice. ## Data freshness & methodology - **Last updated:** 2026-08-11 - **Data as of:** 2026-06 - **What changed most recently:** 2026 hyperscaler AI capex ~$675B, up ~77% YoY from ~$410B. **Methodology (as seeded):** > Two source-backed charts: 2026 AI capex by company ($B) and combined > hyperscaler AI capex versus the US FY2026 defense budget ($B). > > Per-company figures (Amazon ~$200B, Alphabet ~$185B, Meta ~$125B, Microsoft > ~$120B) trace to the cited capex analysis. The aggregate ~$675B is the midpoint > of the $660–690B range for the five largest US cloud/AI providers (incl. Oracle), > up ~77% YoY from ~$410B. Defense ~$1T reflects $838.7B appropriations + $150B > via reconciliation (topline cited near $1.05T). > > CAVEAT: company capex is largely but not exclusively AI. The popular claim that > tech will out-spend the military is a forward projection — in 2026 defense is > still the larger number. Re-verified 2026-07-17. --- _Free tier. The full row-level dataset and per-source detail live on the live page and in the working file. Canonical: . Deepstory Research · https://deepstoryresearch.com_