About AISight
We track how the world's biggest companies are investing in artificial intelligence — from data centers and GPU clusters to R&D and talent.
What We Track
AI CapEx
AI-specific capital expenditure — spending on GPU clusters, AI data centers, custom chips (TPU, Trainium), and AI infrastructure. Where companies disclose AI-specific spending, we use their figures. Otherwise, we estimate based on earnings call guidance ("vast majority for AI") and analyst reports.
Total CapEx
Total capital expenditure from SEC 10-K annual filings. This includes all property, plant, and equipment purchases — not just AI. Sourced directly from XBRL filings via SEC EDGAR.
R&D Spend
Research and development expenses from annual reports. Includes salaries, facilities, and equipment for R&D activities across all business segments — not AI-specific.
Market Cap
Market capitalization updated daily via Financial Modeling Prep API. Represents the total market value of outstanding shares.
Data Sources
- SEC EDGAR — 10-K annual filings via XBRL API for US public companies (CapEx, R&D, Revenue, Employees)
- Financial Modeling Prep (FMP) — Market capitalization, updated daily
- Company Earnings Calls — AI-specific CapEx guidance and management commentary
- Google News RSS — Real-time AI spending news aggregation
- NewsAPI — Supplementary news with article summaries
- Analyst Reports — Bloomberg, CNBC, Reuters, TechCrunch for cross-validation
Confidence Levels
Each data point carries a confidence indicator:
Verified
Confirmed from 10-K filing, earnings release, or authoritative financial data source. Highest reliability.
Estimated (~)
Derived from analyst estimates, management commentary ("vast majority"), or indirect calculation. Shown with ~ prefix and "est." badge.
Unverified
From unconfirmed press reports or limited sources. Treated with caution.
FAQ
How is AI CapEx calculated for companies that don't disclose it?
We use management guidance from earnings calls. For example, when Amazon says "the vast majority of our $131.8B capex is for AWS/AI infrastructure," we estimate ~76% ($100B) as AI CapEx. These estimates are marked with a ~ prefix.
Why do private companies show "funding raised" instead of CapEx?
Private companies (OpenAI, Anthropic, Databricks) don't file public financial statements. We show their total funding raised as a proxy for investment capacity, not actual capital expenditure.
How often is the data updated?
News and market cap data update 3 times daily via automated GitHub Actions. Financial data (CapEx, R&D, Revenue) updates quarterly after earnings season from SEC filings.
Can I use this data in my research?
Yes! You can download the leaderboard as CSV from the homepage. Please cite "AISight (aisight.fyi)" when referencing the data.
How do I report a data error?
Open an issue on our GitHub repository with the company name, incorrect value, and correct source.
Calculated Metrics
We automatically derive these metrics from primary data:
- AI CapEx % of Total — What portion of total capex goes to AI
- CapEx-to-Revenue — How much revenue is reinvested into AI infrastructure
- Capital Intensity — AI CapEx as % of market cap (investment aggressiveness)
- Revenue per Employee — Operational efficiency metric