Why 1.5 Million People Quit ChatGPT…

1.5 million people just boycotted ChatGPT and here's why. So 1.5 million people just left ChatGPT in less than 48 hours. That is not a typo. 1.5 million people, cancellations, deleted accounts, boycott signups in less than two days.

And the story behind why that happened is one of the most important things going on in the AI industry right now. And I'm going to walk you through everything, every number, every fact, every piece of data so that by the end of this, you understand exactly what's happened, what it means for the tools you use, and what it means for your business, your career, and your future with AI. Let's start with the basics. There are three companies that basically run the AI market right now.

OpenAI, which makes ChatGPT, Anthropic, which makes Claude, and Google, which makes Gemini. Those three companies are in a race. And for most of the past three years, OpenAI was so far ahead of that race, it didn't even look close. But this week, something happened that changed the scoreboard.

Here's what went down. The US Department of Defense, which the Trump administration has officially rebranded the Department of War, had a contract with Anthropic. They were using Claude on classified networks. The deal was worth up to $200 million.

And over a period of months, they were negotiating the next phase of that contract. The military wanted what they called all lawful purposes access. That means they wanted to be able to use Anthropic's AI for any case that is currently legal under US law. That would include the collection of Americans' public information in bulk, things like geolocation data, web browsing history, and personal financial information purchased from data brokers.

Anthropic said it needed two things written explicitly into the contract. Number one, the AI could not be used for mass domestic surveillance of American citizens. Number two, the AI could not be used in fully autonomous weapon systems, meaning weapons that make a decision to strike without a human in the loop. The Pentagon maintained that it must be allowed to employ AI systems for any lawful use, which may cross some of Anthropic's red lines.

Talks collapsed. Defense Secretary Pete Hegseth announced he would direct the Defense Department to label Anthropic a supply chain risk to national security, a move usually reserved for foreign adversaries, which would bar any military contractor or supplier from doing business with Anthropic. President Trump ordered every federal agency to immediately cease all of Anthropic's technology with agencies given six months to phase out existing business. Anthropic CEO Dario Amodi responded publicly, threats do not change our position, he said.

We cannot in good conscience accede to their request. Hours later, on that same Friday evening, OpenAI Sam Altman posted on X that his company had reached an agreement with the Department of Defense to deploy its models on their classified network. That sequence of events, Anthropic out, OpenAI in, all in the space of a single Friday, is what triggered everything that followed. ChatGPT uninstalls in the US jumped 295% day over day.

One star reviews of the ChatGPT app surged 775%. Claude downloads rose as much as 51%. Claude hit number one on Apple's US app store for the first time in its history. Anthropic confirmed daily signups had tripled since November, free users had grown over 60% since January, and pay and subscribers had more than doubled, breaking all time records every single day that week.

Prior to this moment, it had been ranked 131st in the US app store on January the 30th. It went from 131 to number one. A boycott movement called QuitGPT emerged and claims over 1.5 million people have taken action either by canceling subscriptions, sharing boycott messages, or signing up at QuitGPT.org. An open letter titled We Will Not Be Divided grew from a couple of hundred signatures on Friday to almost 900 by Monday with nearly 100 signatories from OpenAI itself and close to 800 from Google.

OpenAI's own employees signing a letter in support of the company their CEO just replaced in a government contract. Protesters gathered outside OpenAI's headquarters in San Francisco. Chalk messages appeared on the sidewalk reading things like, is it time to quit? and Orwell warned us.

Sam Altman acknowledged the optics. He said the deal looked opportunistic and sloppy and that the company had genuinely been trying to de-escalate things but shouldn't have rushed. By Monday, Altman had gone back to the Pentagon to renegotiate. OpenAI amended the agreement to include explicit information and language that its AI models shall not be intentionally used for domestic surveillance of US persons and nationals, including through the use of commercially acquired personal or identifiable information.

That's the factual sequence of events. No judgment, just what happened. Now let's zoom out because this story doesn't exist in a vacuum. It's happening on top of a market that was already shifting dramatically before any of this went down.

So these numbers are stunning. ChatGPT's web traffic fell from 86.7% in January 2025 to 64.5% in January 2026. That is a 22% drop in 12 months and that's the overall web traffic share. Google Gemini captured much of the lost, growing from 5.7% to 21.5% over the same period.

In terms of mobile app market share specifically, ChatGPT fell from 69.1% in January 2025 to 45.3% by early 2026. Gemini grew from 14.7% to 25.2%. Grok jumped from 1.6% to 15.2%. The Apptopia data brief from March 2026 confirms a broader picture.

Between August 2025 and February 2026, ChatGPT's share of daily active users among the top seven AI chatbot apps fell from 57% to 42% in the US and from 73% to 57% globally. No single app now has over 50% of daily usage in the US. That is a massive structural shift that's happened before this week's events. And the Pentagon controversy accelerated something that was already well underway.

Meanwhile, Claude's engagement numbers tell a different story than its raw user count. In January, the average time spent per daily user on Claude was the highest of any app at 34.7 minutes. Users aren't just downloading it, they're actually using it. Claude's churn rate, meaning the percentage of users who stop using it each month, fell from 55% in August 2025 to just 36% in February 2026.

That is the largest churn improvement of any app in the entire data set. And on the enterprise set, meaning business customers, not consumers, eight of the Fortune 10 are now Claude customers. The number of customers spending over $100,000 annually on Claude grew 7x in the past year. Over 500 customers now spend more than $1 million annually with Claude.

And that's not a protest bounce, that's a fundamental change in how businesses are choosing their AI tools. Now, let's talk about the financial picture because you can't understand where this is all going without understanding the money. OpenAI's internal projections reported by the information show a $14 billion loss forecast for 2026 alone. The company expects cumulative losses of $150 billion through 2029 before reaching profitability.

OpenAI lost nearly $8 billion in 2025 and only 5% of ChatGPT's 800 million users pay for subscriptions. The company has committed to $1.4 trillion in infrastructure spending over the next eight years. In December 2025, Altman declared a code red internally, delaying advertising and other revenue initiatives to marshal resources toward improving ChatGPT. Google's Gemini topped it on key benchmarks and enterprise market share fell to 27%, whilst Anthropic rose to 40% according to Menlo Ventures.

So the context is here and it's very important. OpenAI is a company dealing with declining market share, massive financial losses, growing competition, and then this week happened on top of all of that. The revenue picture here is real on both sides. Claude Code, Anthropic's AI coding tool, launched in May 2025 and generated over $2.5 billion in annualized revenue by February 2026, more than doubling since the beginning of the year.

Now I want to make sure you understand both sides of this debate clearly because this isn't a simple story. Anthropic pushed for moral boundaries written explicitly into the contract. OpenAI settled for references to existing laws and government policies. The reason OpenAI was able to make a deal when Anthropic could not was by Altman's own admission about approach, not the substance of the restrictions themselves.

OpenAI agreed to the Pentagon's all lawful purposes standard but said it retains full discretion over its own safety stack with strong contractual protections. The Pentagon agreed that publicly available information can only be used for defense and intelligence purposes tied to authorized national security missions. Critics however noted that the published contract excerpt does not give OpenAI an Anthropic-style freestanding right to prohibit otherwise lawful government use. It simply states the Pentagon cannot use OpenAI's tech to break existing laws and policies as they stand today.

The reason Anthropic earned supporters in its fight, including some of OpenAI's own employees, is that they don't believe existing rules are good enough to present and prevent AI-enabled autonomous weapons of mass surveillance. An assumption that federal agencies won't break the law is, in the view of critics, insufficient assurance, particularly for anyone who remembers itself. valence practices exposed by Edward Snowden had initially been deemed legal by internal agencies. That's the technical and legal debate happening at the top levels of this industry right now.

Smart people just what's not in dispute are the numbers and the numbers tell a very clear story. Now I want to zoom out even further because here's the thing about this week that I don't think most people fully process yet. We're watching a market moment where the values of an AI company started to matter to ordinary people in the same way that the values of any other major institution matter. Like a bank or a hospital or a government agency.

For three years people pick their AI tool based on which is the most powerful, which one gives the best answers, which one is easiest to use. This week for the first time at massive scale a huge number of people added a fourth question. Which one do I trust? That's a new variable in this market and it's not going away.

Sociologist Dana Fisher of American University put it plainly. There are lots of examples of failed campaigns like this. The pressure point that might work is if enough people actually use their money to express their political opinions. The structural challenge for any boycott is real.

Jack GPT still has roughly 900 million weekly active users. Even if all 200,000 early quit GPT signups cancelled $20 per month subscriptions, the lost revenue would be roughly $4 million monthly against a company targeting $29.4 billion in 2026 revenue. So in pure financial terms the short-term impact of this week's events is pretty modest for OpenAI. That's the honest number.

The long-term brand and trust impact is much harder to calculate and there are signals beyond the boycott numbers that matter here. Jack GPT's market share dropped from 86% to 65% in a single year before this week. Claude is becoming the dominant tool in coding workflows. Gemini handles research and multimodal tasks for the Google ecosystem.

One of the things driving users toward Claude even before this week is a specific product characteristic. Claude pushes back. It doesn't just agree with you. Multiple users and reviewers have noted that Claude flags flaws in arguments even when not asked to.

For people doing serious work that's not a small thing. So now let's talk about what this means for you. If you're a business owner or a leader who uses AI tools you're making a platform decision right now whether you realize it or not. Every team that builds workflows around one AI tool is making a bet.

A bet on reliability, capability and increasingly a bet on the values and stability of the company behind the tool. This week demonstrated something important. These companies are not interchangeable utilities like electricity. They are organizations run by people making high-stake decisions and those decisions can cause enormous market swings very fast.

For anyone running these tools in production the lesson from this week is clear. Every disruption came from decisions, politics and demand surges that had nothing to do with the AI itself. None of these were model failures, no hallucinations, no technical bugs. Just human decisions that cascaded into market events.

Claude went down multiple times on March 2nd and 3rd. Anthropic attributed the outage to unprecedented demand. The authentication infrastructure buckled under the usage surge. Consumer services including Claude AI, mobile apps and Claude Code went offline for hours per outage.

Even the winner of this week's events had his infrastructure tested. That is a fact worth noting. Scalability is a real constraint and the practical takeaway is this. If you're building serious workflows with AI tools right now you need to be thinking about two things simultaneously.

Number one, which tool is best for your specific use case right now? That answer might be different depending on what you need. Claude excels at long document analysis and precision coding. Gemini dominates for Google workspace users and multi-modal research.

ChatGPT leads in creative writing and has the largest plug-in ecosystem. These are real differences that matter for real work. Second, which companies are building products that match your values and risk tolerance? Because this week proved that the answer to that question can change your entire tool stack overnight.

Not because you choose to change it but because your team members choose to change it or because the platform itself became associated with something that affects how you use it. That is a new kind of business risk and the most sophisticated teams are starting to think about it. Now here's where I want to spend a moment on the bigger picture. The AI market is fragmenting.

That is the most important structural fact of 2026. Three years ago, ChatGPT had 86% of the market. One tool near total dominance and that world is now gone. Today, no single app has over 50% of daily usage in the US.

The AI chatbot market has fractured into multiple serious competitors, each with distinct strengths and distinct user bases. This is actually similar to what happened with social media. Facebook had near total dominance for years, then Instagram grew, then TikTok, then Snapchat, then Twitter, then Substack. And now people use different platforms for different things and no single platform commands the kind of loyalty that Facebook once did.

The same thing is happening with AI and faster. As Apptopia VP of Research Tom Graham put it, ChatGPT built the category but as viable alternatives have scaled, users are naturally diversifying their toolkit. The market could end up looking like streaming where a few major players own the market but multiple players can carve out niches based on product differentiation rather than pure network events. This is exactly right and it has massive implications.

It means that if you're not learning multiple AI tools right now, if you're treating this as a single tool decision, you're already left behind. The professionals winning with AI in 2026 are the ones who know which tool to use for which job, the ones who can switch fluidly, the ones who have built enough depth with the technology that they aren't locked into any single platform. And this is not a moment to wait on. Here's why.

Before the Superbowl on February the 8th, Claude was ranked 42nd on the App Store. Since then it has remained in the top 10 most downloaded apps and is now a top 10 productivity app on iOS in 80 countries, up from just one country at the start of 2025. That's a fundamental shift in the competitive landscape and it happened in just six weeks. Six weeks is nothing.

Six weeks ago, most people had never seriously tried Claude. Now it has more engaged users per session than any other AI app on the market. This is the speed at which this industry moves and the speed is increasing not slowing down. Let me give you a moment here to really absorb what we're describing.

We're talking about an industry where a company can go from 131st on the App Store to number one in a single month, where 1.5 million people can be motivated to cancel subscriptions in 48 hours, where nearly 900 employees at competing companies can sign an open letter in solidarity with their employer's direct competitor, where a CEO publicly admits a deal was opportunistic and sloppy and immediately renegotiates it all within 72 hours. That's not normal business velocity. That's something else entirely. Right now, I want to pause and be direct with you about something.

This is a part of the show where I want to talk to you honestly about what this all means for people like you and me. Whether you're a business owner, a freelancer, a content creator, a developer, a marketer, a manager, whatever role you're in, you're living through a period where the rules of your industry are being rewritten. Not in 10 years, not in five years, right now. The people who understand these tools deeply, who know not just how to use them but why they work the way they do, what their strengths are, what their limitations are, how to combine them into real workflows to save hours, those people have a compounding advantage over everyone else.

And the gap between people who understand this technology and people who don't is getting wider every single month. That's exactly why I built the AI Profit Boardroom. It's a community for people who want to genuinely stay ahead of what's happening in AI. Not just reading headlines but actually learning the tools, building the workflows, and understanding the business implications of every major development like this one.

If you want to be the kind of person who, when a week like this happens, knows immediately what it means for your tools, your team, and your bottom line, that's what we do inside the AI Profit Boardroom. Links in the comments description or just go to the aiprofitboardroom.com and check it out. Now back to the story because there's a few more numbers I want to make sure you have. Let's talk about what Claude actually is right now as a product because a lot of people are hearing the name for the first time this week and they need context.

Claude Opus 4.6, released in early February 2026, took the top spot on the artificial analysis benchmark rankings, beating OpenAI's GPT 5.2 and Google's Gemini. It also ranked number one on ArenaAI where users score AI responses in blind comparative tests across writing, coding, and instruction following. That's the capability picture. The product was already performing at the top of its class before this week's events.

The controversy accelerated adoption of something like that was already technically excellent. Among developers specifically, Claude Code, Anthropx's command line coding agent, had become the most widely used adopted coding agent across startups, enterprises, and research teams entering 2026. Teams that once used it for autocomplete have discovered it functions more like a senior engineer than a line completion tool. And by February 2026, Claude Code generated over $2.5 billion in annualized revenue, more than doubling since the beginning of this year alone.

Now contrast that with the OpenAI financial picture. OpenAI's enterprise market share fell to 27%, whilst Anthropx rose to 40% according to Menlo Ventures data from late 2025. Now let me put that into plain language. In the enterprise market, meaning businesses paying real money for AI tools, Anthropx had already overtaken OpenAI before any of this week's events.

40% versus 27%. That's not a close race. rate is among the fastest of any startup in history, with internal projections showing $14 billion in losses for 2026 alone. The company expects cumulative losses of $115 billion through 2029 before reaching profitability sometime in the 2030s.

One venture capital executive described the situation as the WeWork story on steroids. Sebastian Malabi, senior fellow at the Council on Foreign Relations, wrote in the New York Times that OpenAI could run out of money over the next 18 months, arguing that competitors like Google, Microsoft and Meta can cross-subsidize AI from their existing businesses whilst OpenAI cannot. Now, I want to be completely clear here. OpenAI is not going away tomorrow.

ChatGPT has more than 900 million weekly active users. That's enormous as an installed database. The company has raised hundreds of billions in funding. It has major partnerships with Microsoft, Apple and others.

It is a legitimate and powerful company. The point is not that OpenAI is collapsing. The point is that the dominance that looked permanent in 2022 and 2023 is now genuinely contested. And the events of this week accelerated a contest that was already well underway.

As one developer put it, the comparison analysts are now making isn't to Google. It's to MySpace. MySpace was the dominant social platform. Then Facebook arrived.

MySpace didn't disappear overnight, but within a few years it was totally irrelevant. Not because anything dramatic happened. Because people quietly chose something better and the momentum never came back. Whether that comparison is accurate for OpenAI remains to be seen.

But the fact that analysts are now making it at all tells you something about the moment we're in. Now, let me give you the full picture of the competitive landscape because this matters for how you make decisions about your tools. On the consumer side, ChatGPT has lost US daily active users for four consecutive months and global daily active users for three consecutive months as of the February 2026 data. Google Gemini's monthly active database reached 650 million by October 2025 with 77.9% of interactions occurring on mobile.

So Gemini is now the clear number two globally. Claude had the most dramatic February of any AI app. It's US daily active user market share roughly tripled in a single month, jumping from 1.5% in January to nearly 4% in February. Worldwide it doubled.

On the enterprise side, Anthropic reports serving over 300,000 business customers as of October 2025. Eight of the 1410 are now Claude customers. The company's revenue grew from 1 billion ARR at end 2024 to more than $5 billion by 2025. Anthropic's Sonic 4.5 enabled what companies described as investment grade financial analysis for firms like Nordea and BlackRock.

The company reported 44% faster vulnerability response times for security clients. And on the developer side, the coding tool market has become one of the most important battlegrounds in AI right now. Claude Code, GitHub, Copilot from Microsoft and various other tools are competing for the attention of the world's software developers. And developers matter, not just because they're a big market but because they're the people who build the next wave of AI powered products that everyone else will use.

The company that wins developers in 2026 has an enormous compounding advantage for the next decade. And right now, Claude Code is the most widely adopted coding agent across startups, enterprises and research teams. That position was established before this week's events. This week's events only strengthened it.

As developers who were already leaning toward Claude had one more reason to commit. Now let me give you the forward trajectory because this is ultimately what matters for decisions you make today. The AI market in 2026 is fragmented. That fragmentation is accelerating.

One in five AI users now uses multiple apps suggesting the different tools useful for different tasks, opening the possibility of a market where AI companies carve out niches. The companies that will win in this fragmented market are the ones that establish deep trust with specific user segments, not the ones that try to be everything to everyone. AnthropX's position after this week, stronger brand trust among privacy conscious users, enterprise customers and developers, already the number one tool in enterprise AI by market share. Now has the number one downloaded app in the US market as well.

OpenAI's position, still the largest user base on earth at 900 million weekly active users, still massive revenue but declining market share. A complicated week in terms of public trust and financial losses that require continued enormous investment to sustain. And Google's position, steady growth using distribution advantages. Gemini is built into Android, into search, into Gmail, into docs.

Nearly 75% of Google Cloud customers now use Google's AI products. That distribution moat is real and significant. So here's the practical breakdown for different types of people watching this. If you're a business owner or executive, the AI tool decision for your organization is no longer just a technology decision, it's a vendor relationship decision that includes questions about the values, financial stability and public positioning of the company you're betting on.

The smart move right now is to pick not just one tool, right, don't go all in on one tool. It's really, you should be understanding multiple tools deeply enough that you can switch or diversify as the market continues to shift. If you're a developer or technical professional, Claude Code's position at the top of the coding agent market is worth understanding. If you haven't spent serious time with it, this week is a good moment to do that.

The tool has real product advantages that exist independent of this week's news. Benchmark rankings, context window capabilities and the trust of the developer community built over the past 18 months. If you're a content creator, marketer or knowledge worker, the expansion of Claude's capabilities including deep research, web research integration, artifacts and the ability to handle very long documents makes it genuinely useful for a broader range of daily professional tasks than it was 12 months ago. The market share shift happening right now is partly a political moment and partly a real product story and both are well worth understanding.

And if you're anyone who uses AI for anything, this week proved that the AI you use, the company behind it and the decisions that companies make are things worth paying attention to. Not in a reactive way, not switching tools every time something happens in the AI news, but in a considered informed way. The professionals who will be most effective with AI over the next five years are the ones who understand the landscape clearly, not the ones who follow any single platform loyally regardless of what happens. The AI market has never been more competitive than it is right now, it's never moved faster, it's never been more consequential for ordinary professionals to understand what it is and what's happening and why.

The through line across all of this data is that the AI chatbot market is fragmented. Six months ago chat GPT commanded a near super majority of daily usage, today no single app has a share of over 50% in the US. That sentence, that single sentence is the summary of where we are. The age of one tool dominance is over, especially in AI.

And what comes next is a more complex, more competitive, more fragmented market. One where capability, trust, values, financial stability and fit for purpose all matter simultaneously. And the people who navigate that complexity well, who build real fluency across multiple tools, who understand the business implications of technical developments, who can translate what's happening in the industry into decisions that actually help their work, those people have an enormous advantage right now. Ford's rise to number one is more than a chart position, it reflects a growing segment of users who want AI that's both powerful and principled.

Tools that perform at the frontier without compromising the values of the people building them. Whether that position holds, whether that week's, whether this week's events convert into lasting loyalty, whether OpenAI recovers into its competitive footing again and whether Google's distribution advantage ultimately wins the market, all of that remains to be seen. What is not in question is that the week of February the 28th 2026 was a turning point. In five days, OpenAI signed a military deal, Anthropit refused the same terms and was blacklisted from all federal contracts, Claude hit number one on the app store, then Claude crashed worldwide from unprecedented demand and 700,000 users pledged to cancel their chatuberty subscriptions.

None of these were model failures, no hallucinations, no technical bubs, every disruption came from decisions, politics and demand surges that had nothing to do with AI itself. And that last sentence is the one I want you to carry with you. Every disruption from decisions, politics and demand surges that had nothing to do with the AI itself. The technology itself is advancing, that part is happening regardless, but increasingly the story of AI is not just a technology story, it's a business story, a trust story, a human story about decisions made by real people, real companies that affect hundreds of millions of real people.

And if you want to stay informed, stay ahead and stay ready to act as this story continues to develop, that is exactly what this show is here for. If you found this valuable, share it with one person in your life who uses AI tools and hasn't heard about what happened this week. Forward this episode, send the link, because the more people understand this landscape clearly, the better decisions we can all make. If you want to go deeper, if you want to be part of a community that breaks down every major AI development like this one and translates it into practical apps for your work and your business, check out the AI Profit Boardroom.

The link is in the comments inscription. We break down the news, you can go to the AIProfitBoardroom.com as well. We test the tools, we build the workflows and we make sure that whatever happens next week, you're ready for it. That's everything for today, stay curious, stay informed and I'll see you in the next one.