In early 2026, Anthropic clashed with the US administration over the use of its AI tools for military purposes the company disapproved of. In response, the Department of Defense labelled Anthropic a “supply chain risk” and ended a contract worth up to $200 million.

Is Anthropic merely losing revenue to stand up for its ethical principles, or is prioritizing corporate ethics a smart business strategy?

To find out, Empirical Intelligence and ND Research fielded a choice-based conjoint survey among 314 U.S. consumers to understand what drives people to choose between competing AI subscription plans.

Study Design

We recruited a representative sample of 314 U.S. consumer respondents through online panels. They went through 12 choice tasks where they were asked to select one out of three AI plans with varying features and prices. They could also select none of them. They were told these AI plans were for their personal use and that they would have to pay for them themselves.

The conjoint included a variety of features and prices. There were two options for military use:

  • AI company works with the military
  • AI company does not work with the military

Key Findings

Respondents were considerably more likely to select a plan from a provider that opted out of military use. Estimates from the analysis show that:

  • For an entry-level plan matching the popular ChatGPT Go plan, the preference share increases by 30% when the company opts out of military use.
  • For a premium plan matching the specifications of Anthropic’s Claude Pro plan, the preference share increases by 25% when the company opts out of military use.

Calculating the ethical premium

Using our statistical model to calculate willingness-to-pay (WTP), we quantified the exact dollar value consumers place on companies opting out of military use:

  • Entry-level plan users (like ChatGPT Go users, currently paying $8/month) are willing to pay an extra $2.40/month if the provider does not work with the military.
  • Premium plan users (like Anthropic Claude Pro users, currently paying $20/month) are willing to pay an extra $2.95/month if the provider does not work with the military.
AI Subscription Tier Benchmark Price Preference Increase Ethical Premium (WTP)
Entry-Level Plan (e.g., ChatGPT Go) $8.00 / mo +30% +$2.40 / mo
Premium Plan (e.g., Claude Pro) $20.00 / mo +25% +$2.95 / mo

Strategic Takeaway

These findings show that AI companies don’t necessarily lose revenue by acting ethically. Consumers prefer providers that opt out of military work. They are also willing to pay more for AI plans from these companies.

For AI companies, acting ethically isn’t just a principled stance. It can help protect their bottom line.

Methodology

The survey was run online from March 11 to 20, 2026. Respondents were 18+ and were recruited from a variety of online panels using Cint Exchange. Quotas were used for age, gender, and region to make the results more representative of the US population.