AI Infrastructure Investment Tracker
Tracking hyperscaler capital expenditure commitments against deliverable power capacity. Capital intent ≠ physical delivery — this page audits the gap.
Last Updated: 2026-02-27
Data As Of: CY2024 (10-K) / 2025 guidance (earnings calls) (most recent earnings / 10-K filings)
Sources: SEC 10-K/10-Q filings, earnings call transcripts, press releases — all free, public, clickable.
Total Capex (4 co.)
$245B
2026 guidance
YoY Capex Growth
+67%
aggregate
Grid Adds/yr
~20 GW
US net new
Queue Wait
5.0 yr
avg interconnection
Hyperscaler Capital Expenditure — Public Disclosures
| Company | Total Capex ($B) | AI/DC Share (est.) | YoY Change | Period |
|---|---|---|---|---|
| Microsoft | $55.7B | ~60% | +79% | FY2024 (Jul 2023–Jun 2024) |
| Google (Alphabet) | $52.5B | ~70% | +63% | CY2024 |
| Amazon (AWS) | $77B | ~65% | +60% | CY2024 |
| Meta | $37B | ~80% | +36% | CY2024 |
Capex Commitments vs Deliverable Power
The structural tension at the heart of AISI: hyperscalers can commit capital instantly, but the grid cannot deliver power instantly. The US adds ~20 GW of net new capacity per year, and interconnection queue wait times average 5 years. This creates a multi-year lag between capex announcements and actual power availability.
📊 The capex-power gap
$245B in annual capex commitments vs ~20 GW/yr grid additions. At current queue completion rates (~14%), only a fraction of announced capacity will be operational within the capex planning horizon. This is not a forecast — it is a structural constraint.
Methodology — v0.1-beta
This page aggregates publicly disclosed capital expenditure data from hyperscaler earnings calls, SEC filings, and press releases. AI/DC share estimates are based on management commentary and are approximate.
Data Sources:
- SEC EDGAR — 10-K and 10-Q filings (free, public)
- Earnings call transcripts — via company investor relations pages
- Company press releases and blog posts
- DOE/LBNL — grid capacity context
Known Limitations:
- Companies do not always disclose AI-specific capex; "AI/DC share" is estimated from management commentary.
- Capex guidance may be revised during the fiscal year; this page reflects the most recent public guidance.
- Capital commitments ≠ power delivery — the grid constraint is physical, not financial.
- This is an experimental research tool (v0.1-beta). Not investment advice.
📎 Cite This Data
AhaSignals. (2026). AI Infrastructure Investment Tracker. Retrieved from https://ahasignals.com/ai-infrastructure-investment-tracker/
Frequently Asked Questions
How much are hyperscalers spending on AI infrastructure? ▾
Based on public earnings calls and SEC filings, the four largest hyperscalers (Microsoft, Google, Amazon, Meta) have collectively committed over $200B annually to capital expenditure, with a significant and growing share directed toward AI infrastructure (data centers, GPUs, networking). Exact AI-specific breakdowns are not always disclosed.
Does capex spending guarantee power delivery? ▾
No. Capital commitments and actual power delivery are fundamentally different. The US grid adds approximately 20 GW of net new capacity per year, and interconnection queue wait times average 5 years. Capex announcements represent financial intent, not physical infrastructure delivery. This gap between announced spending and deliverable power is a core AISI stress signal.
Is this investment advice? ▾
No. This page is an independent research audit produced by AhaSignals for educational purposes only. It does not constitute investment advice.
Related Pages
📊 Get Updates
Get weekly updates when the dashboard state changes materially, plus new research on consensus fragility and market divergence. Research-only. Not trade signals.
🔒 No spam. Unsubscribe anytime. 2,000+ researchers and practitioners as of Apr 2026.