Anthropic is making hay while the sun shines.
The AI company’s high-stakes dispute with the Pentagon—in which it refused to allow its product to be used for autonomous weapons and domestic mass surveillance—generated intense mainstream media coverage and a wave of public support, including from many within the AI community. Claude rose to no. 1 on the Apple App Store’s free app rankings on February 28, and on Tuesday it hit number one on a similar ranking in the Google Play store.
The government is effectively banning the use of Anthropic models and tools within government agencies and their suppliers, and has labeled Anthropic as the “woke” AI company. Yet there has been no sweeping social media campaign by MAGA supporters to boycott Claude.
On the contrary, the public appears to have largely sided with Anthropic. A national survey conducted by Morning Consult of 2,000 U.S. adults found that more than half believe the government went too far in its treatment of the company. Two-thirds of Americans, across party lines, said tech companies have a responsibility to set limits on their AI products. Nearly four-fifths said humans, not AI, should make the final decision to use deadly force in war.
Numerous tech influencers on X, including Google’s Jeff Dean, jumped to Anthropic’s defense. Meanwhile, OpenAI appeared opportunistic, moving quickly to sign a less prohibitive and less legally binding contract with the Pentagon. That does not appear to have hurt OpenAI with consumers, though there is no data suggesting it improved the company’s image either. Over the weekend, a viral video showed sidewalks outside Anthropic’s headquarters covered in chalk art and slogans reading “We are proud of you.”
Anthropic’s Super Bowl ad may have helped set this consumer surge in motion. The spot took a shot at rival OpenAI and its ChatGPT product. While it sparked debate on tech Twitter, it appears to have resonated with many consumers. The company reported a significant bump in app signups after the game, and the momentum carried forward.
All of this has cast Anthropic in a favorable light, and the company is capitalizing.
Anthropic recently added memory to Claude’s free plan, so even people who are using Claude for free can start to develop a working relationship with the chatbot. Claude no longer has to start every conversation or project from a clean slate with no context about the user, their interests and preferences, typical projects, productivity challenges, and whatever else. Users can give Claude directions on how they want to work with the AI, what they expect from it. It will remember workflows from past projects.
Anthropic has also made it easier for people to bring that kind of context and preference information over to Claude from competing chatbots, including ChatGPT. The company even gives explicit instructions on how to make the competing chatbot output everything it knows about the user so that the information can be imported into Claude’s memory. Anthropic isn’t trying to capture that user context data so that it can lock users into Claude–the company points out that users can export that data from Claude as easily as they can import it.
Still, users may be inclined to stick with the chatbot that knows them best, especially if that chatbot’s familiarity translates into tangible productivity gains. Indeed, chatbots are steadily moving in that direction, gaining the ability to act autonomously on a user’s behalf to operate computers, generate code, and so forth.
Once users begin experiencing real productivity “wins”, they might come to rely on the tool more frequently, and for more ambitious tasks. At that point, the limits of the free tier can begin to feel restrictive, nudging users to subscribe or upgrade to get the necessary model access and speed. This upgrade pattern may already be visible in the numbers: While Claude’s free users have grown by 60% since the beginning of the year, subscriptions to its Pro and Max plans also more than doubled, the company says.