Sovereign AI 15
Fifteen names with disclosed government AI contracts in the US, UK, France, Germany, Japan, India and the Gulf, weighted by contract value.
Why it won
Sovereign AI is the budget line every Western and Gulf government has been growing for two years. This thesis filters for companies actually winning the contracts, not the broader AI universe, by using disclosed contract values from procurement databases. It is the most precise way to express the theme without buying the Magnificent Seven again.
The prompt
15-stock index of companies winning sovereign AI contracts in the US, UK, France, Germany, Japan, India and the Gulf. Weighted by disclosed annual contract value, max 10% per name.
Methodology
- Universe: any listed company with disclosed sovereign or defence AI contracts above 50m USD per year.
- Geographic scope: US federal and state, UK, France, Germany, Japan, India, UAE, Saudi Arabia.
- Sourcing: USASpending.gov, UK Contracts Finder, EU TED database, company filings.
- Weight by trailing 12m disclosed sovereign AI contract value, capped at 10% per name.
- Quarterly rebalance to capture new contract announcements.
Backtest
Figures are illustrative for editorial purposes. Past performance does not guarantee future results.
Holdings
| Ticker | Name | Weight |
|---|---|---|
| PLTR | Palantir Technologies | 10.0% |
| IBM | International Business Machines | 9.0% |
| ORCL | Oracle | 9.0% |
| BAH | Booz Allen Hamilton | 8.0% |
| LDOS | Leidos Holdings | 8.0% |
| CACI | CACI International | 7.0% |
| SAIC | Science Applications International | 6.0% |
| CRWD | CrowdStrike Holdings | 6.0% |
| NET | Cloudflare | 5.0% |
| S | SentinelOne | 4.0% |
| QQ.L | QinetiQ Group | 5.0% |
| CAP.PA | Capgemini | 6.0% |
| ATO.PA | Atos SE | 4.0% |
| 9613.T | NTT Data Group | 7.0% |
| TCS.NS | Tata Consultancy Services | 6.0% |
Risks
- Palantir concentration risk is structural to the theme.
- Contract disclosure data is lumpy, so quarterly rebalances will see large weight swings.
- Sovereign AI budgets are subject to political rotation and a single administration change can re-price the basket.
Earlier weeks
Thirty global compounders with ROIC above 20% and zero exposure to the Magnificent Seven, on a 5-year backtest that quietly clears SPY.
Twenty defence and aerospace names with no exposure to cluster munitions, nuclear delivery systems or anti-personnel mines, weighted by 3-year backlog growth.
The 25-year dividend-growth list with tobacco names dropped and the savings redistributed by FCF coverage, not yield.
