How the answers frame RTX
This is a single snapshot taken on 2026-09-29. It covers ChatGPT, Claude, Gemini, Perplexity and Grok, and 250 benchmark questions per assistant that never named a brand, for 1,250 answers in total. RTX was named in 393 of them. By assistant that was 119 of 250 for Grok, 79 each for ChatGPT and Gemini, 59 for Claude and 57 for Perplexity. In this snapshot Boeing, Lockheed Martin and Northrop Grumman were each named in more answers than RTX.
The dominant frame is membership in the group of large U.S. primes. Claude's answer on trust described it as a set: "The Big 5 contractors—Lockheed Martin, RTX, Northrop Grumman, Boeing, and General Dynamics—control most US defense spending across aircraft, missiles, and naval systems." Perplexity and Grok put RTX among the main missile defense contenders. Claude's revenue answer listed RTX second among global defense contractors by defense revenue.
Few answers credited RTX with a named quality. In the coded credits, "Not Substantive" dominated (11), followed by Financial Performance (4), Safety & Compliance (2) and Price & Value (1). RTX did better on broad questions than on specific ones. Visibility by question intent was highest for leadership (61%) and government contracts (56%). It was lowest for safety (10%), value (11%) and service (14%).
Some answers put RTX in compliance and cost contexts. Claude's export-control answer grouped RTX and GE Aerospace with Boeing under penalties "from $36 million to $200 million." ChatGPT's shareholder-suit answer said RTX disclosed notice of multiple shareholder derivative lawsuits. Perplexity's budget answer cited an analysis in which Raytheon performed relatively poorly on cost efficiency. These are how the answers described it, not PeakMetrics' assessment.
The Big 5 contractors—Lockheed Martin, RTX, Northrop Grumman, Boeing, and General Dynamics—control most US defense spending across aircraft, missiles, and naval systems.
What the press is saying
The PeakMetrics news index recorded 7,522 mentions of RTX over the last 30 days, 3,634 of them with the name in the headline. That is roughly the same volume as Northrop Grumman (7,170) and far below Boeing (47,588) and SpaceX (51,777).
The recent headlines supplied do not describe the defense company. They cover an Nvidia RTX 50-series memory supply story, discounted RTX 50-series gaming PCs, PC specs for a video game, an AI-training blog post and a crypto token presale using the RTX label. The name is shared with other products and tickers, so a large part of the RTX news count looks unrelated to the defense business.
Raytheon and General Dynamics performed relatively poorly on cost efficiency in the programs studied.
The gap
The press number and the AI story are not measuring the same thing. In the news index, RTX looks like a heavily covered name, but the sampled headlines are consumer-tech and crypto items. In the assistant answers, RTX is a prime contractor, defined by scale, missile defense work and government contracts.
The two sets of peers also differ. Boeing and SpaceX draw far more news than RTX, and RTX draws about as much as Northrop Grumman. In this snapshot RTX's answer volume sits between Northrop Grumman and Airbus. Press volume does not predict how often the assistants name a defense company. It also says little about what they credit it for.
What to watch
First, the answers name RTX often but rarely attach a specific strength. A comms team could check which proof points, such as delivery, safety or value, the assistants have available to cite. Safety and value questions are where visibility was lowest in this snapshot.
Second, the cited material in compliance, litigation and cost answers comes from filings and third-party studies. Those are the sources the assistants drew on when RTX appeared in those contexts.
Third, raw news counts for RTX include a lot of non-defense content, so media monitoring needs filters to separate the company from GPU and crypto coverage. These are single runs of live answers, and results will vary between runs. Treat this as a baseline to re-measure, not a ranking.
