AI Data Center Power by Company
Independent audit of hyperscaler electricity consumption using only publicly disclosed data. Sustainability reports, SEC filings, and earnings calls — no paywalled estimates.
Last Updated: 2026-02-27
Data As Of: CY2024 (Google), CY2023 (Meta), FY2024 (Microsoft, Amazon) (most recent sustainability reports)
Sources: Corporate sustainability reports, SEC 10-K filings, earnings transcripts — all free, public, clickable.
Total (4 Hyperscalers)
~100 TWh
2023–2024
YoY Growth
+~25%
aggregate
Largest Consumer
30.8 TWh
% of US DC Total
~55%
est. share
Hyperscaler Electricity Consumption — Public Disclosures
All figures below are sourced from publicly available corporate sustainability reports and SEC filings. Methodologies differ across companies — direct comparison requires caution.
| Company | Total Energy (TWh) | YoY Change | Renewable % | Reporting Period |
|---|---|---|---|---|
| Microsoft | Not disclosed | +N/A% | 100% | FY2024 (Jul 2023–Jun 2024) |
| Google (Alphabet) | 30.8 | +27% | 100% | CY2024 |
| Amazon (AWS) | Not disclosed | +N/A% | 100% | CY2024 |
| Meta | 15.0 | +34% | 100% | CY2023 |
⚠️ Comparability caveat
Each company uses different reporting boundaries, PUE assumptions, and renewable energy accounting methods (market-based vs location-based). Direct comparison of absolute TWh figures requires understanding these methodological differences. See Methodology section below.
Data Freshness
| Source | Frequency | Data As Of | Status |
|---|---|---|---|
| Microsoft Sustainability Report | Annual | Jun 2025 | PARTIAL |
| Google Environmental Report | Annual | Jun 2025 | CURRENT |
| Amazon Sustainability Report | Annual | Jul 2025 | PARTIAL |
| Meta Sustainability Report | Annual | Aug 2024 | CURRENT |
Methodology — v0.1-beta
This page aggregates publicly disclosed electricity consumption data from the four largest hyperscalers. It does not estimate or model undisclosed consumption — only reported figures from official corporate publications are used.
Data Sources (all free, public, clickable):
- Microsoft Environmental Sustainability Report (annual)
- Google Environmental Report (annual)
- Amazon Sustainability Report (annual)
- Meta Sustainability Report (annual)
- SEC 10-K filings (supplementary)
Known Limitations:
- Reporting boundaries differ: some companies include all facilities, others only owned data centers.
- PUE (Power Usage Effectiveness) assumptions vary; some report IT load only, others total facility load.
- Renewable energy accounting uses market-based instruments (RECs/PPAs) which may not reflect physical grid mix.
- Sustainability reports typically lag 6–12 months behind the reporting period.
- This is an experimental research tool (v0.1-beta). Not investment advice.
📎 Cite This Data
APA:
AhaSignals. (2026). AI Data Center Power by Company. Retrieved from https://ahasignals.com/ai-data-center-power-by-company/
Frequently Asked Questions
How much electricity do hyperscalers consume? ▾
Based on publicly disclosed sustainability reports, the four largest hyperscalers (Microsoft, Google, Amazon, Meta) collectively consumed an estimated ~100 TWh in their most recent reporting years (2023–2024). Individual figures vary significantly and are disclosed with different methodologies, making direct comparison challenging. Google (30.8 TWh) and Meta (15.0 TWh) disclose exact figures; Microsoft and Amazon do not publicly report total electricity consumption.
Where does hyperscaler power consumption data come from? ▾
All data on this page comes from publicly available sources: corporate sustainability reports (ESG filings), SEC 10-K filings, earnings call transcripts, and press releases. No paywalled research or proprietary estimates are used.
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.