U.S. consumers uninstalled ChatGPT's mobile app at 295% above normal rates on Saturday, February 28, according to market intelligence firm Sensor Tower. The spike followed OpenAI's announcement that it had reached a deal to deploy its AI models on the Pentagon's classified network. ChatGPT's typical day-over-day uninstall rate over the past 30 days sat at 9%.
The deal landed hours after the Trump administration blacklisted Anthropic, OpenAI's chief rival, for refusing to let the Department of War use its AI technology without restrictions on mass domestic surveillance and autonomous weapons. Anthropic's Claude climbed to No. 1 on Apple's U.S. App Store that same Saturday, a position it still held on Monday, March 2. Consumers delivered their verdict before Monday.
What Changed
- ChatGPT uninstalls surged 295% day-over-day after OpenAI signed a Pentagon AI deal, per Sensor Tower.
- Claude hit No. 1 on the U.S. App Store; three analytics firms confirmed the consumer shift.
- Altman admitted the deal looked 'opportunistic and sloppy' and amended it to ban domestic surveillance.
- Anthropic lost government contracts but gained users, paid subscribers, and cultural momentum.
The download data tells one story
Sensor Tower's numbers were brutal. Claude downloads jumped 37% day-over-day on Friday, February 27, when Anthropic's standoff with Defense Secretary Pete Hegseth became public. By Saturday, that figure hit 51%. ChatGPT downloads, which had grown 14% on Friday before the deal went public, dropped 13% on Saturday and slid another 5% on Sunday.
App Store reviews hit harder still. One-star ratings for ChatGPT surged 775% on Saturday. They grew another 100% day-over-day on Sunday. Five-star reviews dropped by half over the same period. Not a blip. A judgment.
A second analytics firm, Appfigures, confirmed the pattern. Its data showed Claude's total daily U.S. downloads surpassed ChatGPT's for the first time on Saturday, with Appfigures pegging Claude's day-over-day growth at 88%, higher than Sensor Tower's estimate. Claude now ranks as the No. 1 free iPhone app in six countries beyond the U.S., including Canada, Germany, and Switzerland.
Three different analytics firms measured the same weekend and landed on the same conclusion. Every arrow pointed away from ChatGPT.
Anthropic said free users had increased more than 60% since January, with daily sign-ups tripling since November and breaking internal records every day last week. Paid subscribers have more than doubled this year. Before Anthropic ran its Super Bowl ads in early February, Claude sat at No. 42 on the App Store. It entered the top 10 and stayed there.
Similarweb, a third provider, estimated Claude's U.S. downloads over the past week at roughly 20 times January levels, though it cautioned that factors beyond the Pentagon dispute may have contributed.
Altman calls his own deal "opportunistic and sloppy"
Sam Altman's public admission arrived Monday. In an internal memo posted to X, the OpenAI CEO said the company was amending its Pentagon contract to explicitly prohibit domestic surveillance of U.S. citizens. The new language invokes the Fourth Amendment and two Cold War-era surveillance statutes.
Altman's own words were blunt. "We shouldn't have rushed to get this out on Friday. The issues are super complex, and demand clear communication." He added that the company was "genuinely trying to de-escalate things," but conceded it "just looked opportunistic and sloppy."
The word "sloppy" landed hard for a company that positioned itself as the responsible steward of artificial general intelligence. OpenAI had watched Anthropic refuse the Pentagon's terms, watched the administration retaliate by labeling Anthropic a "supply-chain risk to national security," and then moved in with its own agreement before the weekend started. Employees noticed. OpenAI looked cornered. Chalk appeared on the sidewalk outside OpenAI's headquarters that weekend. "NO TO MASS SURVEILLANCE." A second scrawl told employees to do the right thing.
Stay ahead of the curve
Strategic AI news from San Francisco. No hype, no "AI will change everything" throat clearing. Just what moved, who won, and why it matters. Daily at 6am PST.
No spam. Unsubscribe anytime.
Altman also claimed the Pentagon had confirmed its intelligence agencies, including the NSA, would not use OpenAI's services without a separate contract modification. Altman also told employees he'd sooner accept a jail sentence than carry out unconstitutional surveillance. We'll see. Three days earlier, the same CEO couldn't get a press release right.
From Reddit threads to Katy Perry
The consumer backlash organized fast. A website called QuitGPT launched a boycott, claiming over 1.5 million participants by Monday, though it offered no citation for the figure. The Instagram account "quitGPT" gained roughly 10,000 followers in days. A Reddit post urging users to "Cancel and Delete ChatGPT" pulled 30,000 upvotes.
People on social media pointed to OpenAI president Greg Brockman's previous $25 million donation to a pro-Trump super PAC as additional motivation for switching. Pop singer Katy Perry posted a screenshot of her Claude Pro subscription with a heart emoji superimposed over it.
Anthropic moved quickly to capitalize. The company launched a memory import tool that lets users transfer preferences and conversation context from ChatGPT to Claude in a few steps. "With one copy-paste, Claude updates its memory and picks up right where you left off," the company's website stated.
Then the servers buckled. On March 2, Claude experienced a roughly three-hour outage as thousands of new users flooded the platform, according to Downdetector reports.
What the Pentagon deal revealed
OpenAI still has over 900 million weekly ChatGPT users. A weekend of angry uninstalls will not collapse that base. But the episode exposed something the company's leadership had not priced in: consumer AI products carry political risk now, and the cost of a misstep shows up in real time on the App Store charts.
Anthropic lost its government contracts. The Treasury Department, Federal Housing Finance Agency, Fannie Mae, and Freddie Mac all announced they would end Anthropic agreements on Monday. Defense Undersecretary Emil Michael called Dario Amodei a "liar." President Trump labeled the company "Leftwing nut jobs."
And yet Anthropic emerged emboldened. By every consumer metric available, the company won the week. Anthropic refused the Pentagon's terms and gained users. OpenAI accepted them, lost users, amended the deal, and had its own CEO call the result sloppy. You don't need a focus group to read that signal.
Whether the App Store bounce lasts is a different question. Download spikes driven by political moments tend to fade. The 20x increase Similarweb measured includes momentum from Anthropic's Super Bowl campaign and the growing popularity of Claude Code in developer circles. Separating political sympathy from product preference will take months of data that nobody has yet.
What won't fade is the verdict itself. AI companies now operate in a market where a single government contract can trigger a consumer revolt measured in hundreds of percentage points. Enterprise deals that once stayed in boardrooms now get judged on App Store charts. The chalk is still on the sidewalk outside OpenAI's office on Mission Street.
Frequently Asked Questions
Why did ChatGPT uninstalls spike 295%?
OpenAI announced a deal to deploy AI on the Pentagon's classified network on February 28, hours after the Trump administration blacklisted rival Anthropic for refusing unrestricted military use. Sensor Tower data showed ChatGPT's typical 9% daily uninstall rate jumped to 295% above normal.
What changes did OpenAI make to the Pentagon deal?
On Monday, OpenAI amended the contract to explicitly prohibit domestic surveillance of U.S. citizens, invoking the Fourth Amendment and Cold War-era surveillance statutes. Altman said the NSA would not use OpenAI services without a separate contract modification.
How much did Claude downloads increase?
Claude downloads jumped 37% on Friday and 51% on Saturday, per Sensor Tower. Appfigures put Saturday growth at 88%. Similarweb estimated downloads at roughly 20 times January levels. Claude reached No. 1 on the U.S. App Store and topped charts in six other countries.
Why was Anthropic blacklisted by the Pentagon?
Anthropic refused to let the Pentagon use its AI for 'all lawful purposes,' including mass surveillance and autonomous weapons. Defense Secretary Pete Hegseth designated the company a 'supply-chain risk to national security,' and Trump ordered federal agencies to phase out Anthropic technology within six months.
Will Claude stay at No. 1 on the App Store?
Likely not permanently. Download spikes driven by political moments tend to fade. But Anthropic's growth predates the dispute: free users are up 60% since January and paid subscribers have more than doubled this year. The company also launched a memory import tool for ChatGPT switchers.
