Two defense technology companies are being cited inside AI assistants at a rate that exceeds the five largest U.S. defense primes combined, according to a benchmark published this week.
The Defense & Aerospace AI Visibility Index 2026 — produced by 5W and built on 28,400 prompts across ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews — found that Anduril Industries (19.8%) and Palantir Technologies (15.2%) together accounted for 35.0% of AI Citation Share across the U.S. defense and aerospace category. Lockheed Martin, Northrop Grumman, RTX, Boeing Defense, and General Dynamics combined accounted for 21.1%.
The two defense tech firms operate at roughly $3.5 billion in combined annual revenue. The five primes operate at approximately $247 billion. Revenue, in this dataset, did not predict citation density. The relationship trended inverse.
SpaceX finished third overall at 12.7% Citation Share — distributed across pure space queries (where it took roughly 60% of subset share), defense queries tied to Starshield and Department of Defense contracts, and adjacent technology queries. No other company in the test set exhibited comparable cross-category presence.
Shield AI placed fourth at 6.4%. Lockheed Martin placed fifth at 5.9%.
Defense trade press dominated the citation surface. Defense News (10.8%), Breaking Defense (8.9%), The War Zone (7.2%), War on the Rocks (5.4%), and Defense One (4.6%) collectively accounted for 36.9% of cited sources. Wikipedia anchored the distribution at 20.7%. Reuters and Bloomberg defense coverage combined for 6.8%. The New York Times, Wall Street Journal, and Financial Times defense desks combined for 5.9%. Substack-hosted war journalism appeared in 4.4% of responses — larger than Defense One and roughly equal to Bloomberg's defense coverage.
The pattern inverts what prior AI Visibility Index volumes recorded in financial services and venture capital, where mainstream financial and business press dominated citation share.
Founder visibility correlated with company citation density. Companies whose founders maintained public writing or long-form podcast activity during the testing window — Palmer Luckey at Anduril, Alex Karp at Palantir, Elon Musk at SpaceX, Brandon Tseng at Shield AI — were associated with higher observed Citation Share. The benchmark notes that the association does not establish causation. No legacy prime in the test set exhibited a comparable level of named-founder content output during the measurement period.
Reputational events persisted in retrieval across both waves. References to the Boeing 737 MAX episode — a commercial aviation event — appeared in 31% of Boeing Defense responses. References to F-35 cost overruns appeared in 28% of Lockheed Martin responses. References to Palantir contractor and contract-related discussions appeared in 19% of Palantir responses. The benchmark suggests reputational events embedded in high-citation surfaces like Wikipedia and press archives may not decay with the news cycle.
Engine-level patterns varied. ChatGPT produced higher Palantir share (17.8%) than Anduril share (17.2%) — a pattern the benchmark associates with Wikipedia-weighted retrieval and longer-established entity reference density. Gemini and Perplexity produced higher Anduril share (22.1% and 23.8%) — a pattern associated with retrieval architectures weighting recency and recent social signals. Claude produced the most uniform distribution across the top three companies. Google AI Overviews exhibited the highest within-engine variance.
The Index defines Citation Share as a composite of mentions, recommendations, and source citations — weighted equally per retrieved response. The benchmark explicitly notes the metric is a proxy for model surface frequency, not for market share, contract-award likelihood, or operational performance.
The full dataset, methodology, and engine-level results are published at 5wpr.com/ai-visibility-index/defense-aerospace.
