Claude Rises to No. 1 on the App Store After ChatGPT Uninstall Surge

On Saturday morning, Ultraman shared a screenshot of an internal memo on X.

He wrote it to OpenAI employees on Thursday night, saying the company was in talks with the Pentagon and that he hoped to help "de-escalate the situation." He shared the letter with a few lines of context, aiming to publicly clarify what had happened over the past few days.

When he posted this tweet, Claude had already risen to number one on the free chart of the U.S. App Store. The day before, ChatGPT had held that position.

Sensor Tower’s data recorded what happened in the following hours: On Saturday alone, ChatGPT’s uninstallations in the U.S. surged by 295% compared to average days, and 1-star reviews increased by 775%. Meanwhile, Claude’s daily downloads rose by 51%. A wave of “Cancel ChatGPT” posts appeared on Reddit, with users sharing screenshots of their unsubscribes, and some comments reading “fastest install of my life.” A website called QuitGPT.org went live, claiming that 1.5 million people had taken action.

On Monday, Claude experienced a major outage due to an overwhelming influx of users—the company, designated as a “supply chain security risk” by the federal government, faced server strain from the surge in traffic.

On the same day the uninstallation wave intensified, Anthropic launched its memory migration tool.

The feature itself is simple. Users copy a prompt into ChatGPT, have it output all stored memories and preferences, then paste it into Claude, which imports it with one click and picks up right where you left off in ChatGPT. The website copy says just one sentence: “switch to Claude without starting over”.

The timing of this tool is its most critical attribute.

OpenAI's own data shows that by mid-2025, over 70% of ChatGPT usage scenarios are non-work-related, including everyday Q&A, writing, entertainment, and information lookup. It is many people's first encounter with AI, embedded into daily life through its extensive plugin ecosystem, Voice Mode, and deeply integrated third-party applications. For these users, the switching cost isn't just "downloading a new app"—it's making a new AI, unfamiliar with them, learn who they are from scratch. The accumulation of memories has been the strongest reason to stay.

Anthropic's own research data shows that Claude's usage is highly concentrated. Programming and mathematical tasks account for 34%, making them the single largest category, while education and research were the fastest-growing areas over the past year. The core users are developers, researchers, and heavy writers—individuals who are more rational and more likely to switch tools based on a clear value judgment, as long as the migration cost is low enough.

The memory migration tool has minimized this cost to the lowest possible level. Meanwhile, Anthropic has announced that the memory feature is now fully available to all free users, a feature previously exclusive to paid subscribers.

However, a significant portion of these new users are not the original target audience for Claude.

Based on feedback from social media, many ordinary users who migrated from ChatGPT often react upon first using Claude with: “It’s different.” Some find Claude’s responses more thoughtful, willing to push back rather than simply agreeing with everything. Others notice its writing is cleaner, but it doesn’t generate images or offer interactive experiences like Voice Mode.

Someone initially sought a “more obedient alternative to ChatGPT,” but found that Claude has a stronger personality and requires time to adapt. A migration guide from TechRadar, titled “I Wish Someone Had Told Me These Things,” was widely shared over the past two days, with its core message being: the usage logic of Claude and ChatGPT is fundamentally different—the former is more like a work partner with opinions, while the latter is more like a versatile assistant.

This difference, originally a reflection of each product’s positioning, was unexpectedly amplified by this event. Users, driven by moral convictions, flocked to Claude—only to encounter a product different from their expectations: a more discerning, boundary-conscious AI. What might have been a reason to leave became, at this unique moment, a reason to stay: when you believe in a company’s stance, you’re more likely to accept its product’s logic.

Days after launch, Anthropic released data showing that free active users increased by over 60% compared to January, and daily new registrations quadrupled. Claude experienced an outage due to excessive traffic, with thousands of users reporting inability to log in; the issue was resolved within hours.

Anthropic is the first commercial company to deploy an AI model on the U.S. military’s classified network, with the partnership facilitated through Palantir and a contract valued at approximately $200 million. However, over the past several months, the relationship between the two parties has steadily deteriorated. The core dispute centers on a clause: the Pentagon demanded that the AI model be available for “all lawful uses” without restrictions. Anthropic, however, insisted on including two exceptions: the model must not be used for mass surveillance of U.S. citizens or for fully autonomous weapons systems.

Around February 20, it was reported that an Anthropic executive raised questions with partner Palantir regarding the use of Claude in the U.S. military’s January operation to capture Venezuelan President Maduro, prompting strong dissatisfaction from the military. On Thursday, the Pentagon issued a final deadline, requiring Dario Amodei to respond by 5 p.m. that day.

Amodei issued a statement before the deadline stating that the company could not accept the current terms, “not because we oppose military use, but because in rare cases, we believe AI could undermine rather than defend democratic values.” Trump immediately announced a six-month full suspension of Anthropic products across federal agencies, and Hegseth classified them as a “supply chain security risk”—a label typically reserved for foreign adversary companies. The contract was terminated.

The vacant position was quickly filled. Later that same day, OpenAI announced a contract with the Pentagon. In his internal memo on Thursday, Altman maintained a clear stance, writing that this was already “an industry-wide issue,” stating that OpenAI and Anthropic share the same “red lines”: opposition to mass surveillance and autonomous weapons. On Friday, the agreement was finalized, with the model to be deployed on military classified networks, restricted to cloud-only operation, with engineers stationed to oversee it, and the same two restrictions explicitly written into the contract.

Later, Ultraman opened a Q&A session on X, answering questions for several hours. Someone asked him: “Why did the Pentagon accept OpenAI but ban Anthropic?” His response was: “Anthropic seems more focused on specific prohibitions in contracts rather than citing applicable laws, while we are satisfied with citing the laws.”

This statement speaks to a methodological difference, but it opens the door to the real controversy at hand.

The key issue in Anthropic's breakdown was the phrase the Pentagon insisted on including: AI systems could be used for "all lawful purposes." Anthropic refused because, in the context of national security, this phrase does not constitute a fixed boundary. Current laws have not yet caught up with AI's capabilities, and the scope of what is "lawful" would be determined by the government's own interpretation. OpenAI signed onto this phrase, while claiming that similar protections were discussed in the contract.

Legal experts then analyzed the contract terms publicly disclosed by OpenAI and identified two specific wording issues.

The monitoring clause states that the system must not be used for "unconstrained" monitoring of U.S. citizens' private information. Samir Jain, Vice President of Policy at the Center for Democracy & Technology, points out that this wording implies that "constrained" monitoring versions are permitted. Under the current legal framework, the government can legally purchase citizens' location records, browsing histories, and financial data from data brokers and use AI to analyze this information—technically not constituting "illegal surveillance." Amodei cited this exact example in a subsequent interview with CBS.

The weapons clause states that the system shall not be used for autonomous weapons "in cases where laws, regulations, or departmental policies require human control." This qualifying phrase means the restriction only takes effect if other provisions already mandate human control, and its enforceability is entirely dependent on existing policies. The Pentagon has the authority to modify its internal policies at any time. Legal scholar Charles Bullock wrote on X that the weapons clause in the contract relies on DoD Directive 3000.09, which requires commanders to retain "an appropriate level of human judgment," and this "appropriate level" is a standard open to flexible interpretation.

OpenAI's response to these concerns is that the model can only run on the cloud, which architecturally excludes the possibility of direct integration into weapon systems. The contract also specifies concrete legal grounds, which are more binding than mere textual prohibitions, as laws represent established, tested frameworks. Altman himself acknowledged in his response: “If we ever need to fight this battle, we will, but it clearly exposes us to certain risks.”

This isn’t a matter of one company willing to compromise while the other stands by its principles—it’s a matter of two fundamentally different security philosophies. OpenAI’s bottom line is: I won’t do anything illegal. Anthropic’s bottom line is: I won’t do anything I believe is wrong, even if the law hasn’t yet banned it.

This divide has also left cracks within OpenAI. Last week, multiple OpenAI employees signed an open letter supporting Anthropic’s position and opposing its classification as a supply chain risk. Alignment researcher Leo Gao publicly questioned whether the company’s contracts provided adequate protection. Chalked graffiti criticizing the company appeared on the sidewalk outside OpenAI’s San Francisco office, while supportive messages were left outside Anthropic’s office. Altman’s hours-long Q&A on X was, to a significant extent, directed at his own company’s employees who had originally sided with Anthropic.

For years, Anthropic has framed its safety mission as “preventing civilization-level risks,” equating the potential threats of frontier AI with nuclear weapons and positioning itself as the gatekeeper on this frontline. This narrative is central to its brand and the way it gains trust in capital markets.

Tech commentator Packy McCormick cited Ben Thompson’s concept of the “Hype Tax” during the escalation of this issue: if you build your influence through extreme narratives, you will be charged for it when those narratives confront real power. If you compare AI technology to nuclear weapons, the government will treat you as if you’re dealing with nuclear weapons.

Anthropic paid a price for this narrative: losing a contract, being labeled a security risk, being named by the president, and having all its products required to be removed from federal systems within six months.

But over the same weekend, the same narrative produced a completely opposite effect in another dimension.

Ordinary users don’t see contract terms, legal interpretations, or debates about security philosophy. They see: one company says no and gets kicked out by the government; another says yes and secures the contract. They make their choice using their own framework—295% uninstalls, #1 on the App Store, downed servers.

This is a rare collective moral stance by consumers in the history of the AI industry.

Anthropic did not spend a single dollar on public relations for this matter. Amodei’s statement was restrained, making no appeal for user support, naming no competitors, and portraying no self as a martyr. Yet the outcome occurred anyway.

There is a notable detail here: the event that drove users to Claude was essentially OpenAI doing something entirely reasonable from a business standpoint—securing an agreement while competitors were banned and contracts were voided, and claiming that the same protection terms had been discussed. Altman also explicitly stated that he did this in part to help de-escalate the situation and prevent further harm to Anthropic.

Regardless of the motivations, the result was that OpenAI secured the contract, while Anthropic’s user base grew. Both sides incurred costs and reaped benefits—just measured in different units.

There is one more thing worth mentioning here.

Anthropic lost a Pentagon contract worth approximately $200 million.

Anthropic's current annual revenue is $14 billion, with a target of reaching $26 billion by 2026.

Anthropic completed its $30 billion Series E funding last month, with a valuation of $380 billion.

This calculation isn’t hard now. But another question remains unanswered: When AI is truly used at scale for military decision-making, will the “technical guardrails” outlined in contracts and the engineers deployed actually work—whether those required by OpenAI or Anthropic?

This issue is not covered in any publicly available contract.

Click to learn about the open positions at BlockBeats

Welcome to the official BlockBeats community:

Telegram subscription group: https://t.me/theblockbeats

Telegram group: https://t.me/BlockBeats_App

Official Twitter account: https://twitter.com/BlockBeatsAsia