TL;DR
Last week OpenAI announced ads in ChatGPT’s free tier. Within hours, Claude launched a “No Ads, Ever” campaign. Twitter turned into a roast session. Tech influencers dunked. Users threatened to switch.
“ChatGPT sold out.” “Claude is the good guys now.” “This is the beginning of the end.”
The thread I kept seeing: OpenAI betrayed users for profit while Claude stayed true to their values.
Except I’ve watched this exact movie play out twice before.
Let’s Talk Numbers
ChatGPT has 900 million weekly active users. 58% are on the free tier. That’s 520 million people using ChatGPT without paying anything. Claude has about 20-30 million monthly active users.
ChatGPT serves 30x more people. Different scale entirely.
Here’s where it gets interesting: OpenAI is burning around $9 billion in 2025, with projected losses of $14 billion in 2026. They won’t hit profitability until 2029.
Meanwhile, Claude is also unprofitable. They’ve raised over $37 billion total and are seeking another $20 billion at a $350 billion valuation.
Different user bases though.
User Base Comparison
| Metric | ChatGPT Users | Claude Users |
|---|---|---|
| Personal use (homework, recipes, questions) |
70% | 16% |
| Work-related | 30% | 17% (outside coding) |
| Coding & mathematical work | Minority | 34% of all tasks |
| Demographics | Ages 25-34 biggest group Gender split ~50/50 |
77% male, 52% ages 18-24 |
| Revenue source | Mixed consumer + enterprise | 80% from enterprise APIs |
| User profile | Mainstream: your mom, college students | Developers who read API docs for fun |
Two companies at wildly different scales with different business models.
The Product Adoption Curve
There’s a framework that explains this pattern.
When a new technology launches, adoption happens in stages:
Innovators and Early Adopters make up about 16% of the total market. These are tech enthusiasts. People who’ll pay premium prices to try new things. They want the cutting edge.
Early Majority and Late Majority make up about 68% of the market. These are mainstream users. Price sensitive. They want it to work reliably and they want it cheap or free.
💡 Critical Insight
You can monetize the 16% with premium subscriptions. They’ll pay $20-100/month without thinking twice. But the 68%? They want it free. And if you try to charge them, they’ll just leave for whoever offers it free.
This creates a fundamental split in business models:
Serving the 16%: Premium subscriptions work. Enterprise contracts work. Your costs are manageable because you’re not serving hundreds of millions of users. Examples: Superhuman ($30/month email), Roam Research ($15/month notes), most developer tools.
Serving the 68%: You need freemium with ads. Free tier to acquire users, ads to monetize them, premium tier to convert the ones willing to pay. Your costs are massive because you’re serving hundreds of millions. Examples: Spotify, YouTube, Instagram, Reddit.
The transition from 16% to 68% is where every platform makes The Choice. And the math doesn’t care about your marketing promises.
Claude right now serves the 16%. Their user base is 77% male, 52% ages 18-24, heavily developer-focused. 34% of all tasks are coding and mathematical work. Only 16% use it for personal tasks.
ChatGPT hit the mainstream. 900 million weekly users means they’re deep into the 68%. 70% use it for personal tasks. Your mom uses it. College students use it for homework. Random people who’ve never thought about AI in their lives are using it.
The 68% won’t pay $20/month for an AI chatbot. They want it free or they’ll just not use it.
Instagram’s Journey
April 2012: Facebook acquires Instagram for $1 billion. The app has 30 million users. Zero revenue.
Mark Zuckerberg posts publicly: “We need to be mindful about keeping and building on Instagram’s strengths and features rather than just trying to integrate everything into Facebook.”
Translation: we won’t ruin this with ads immediately. Everyone relaxes. Instagram stays ad-free for over a year.
November 2013: Instagram announces ads will start appearing in feeds.
The backlash is immediate and loud. Users flood tech blogs with comments about how Instagram sold out. Articles predict mass exodus. Twitter fills with people threatening to delete the app.
Instagram proceeds anyway. They roll out “carefully curated brand posts” from a handful of major brands. They promise to do ads differently than Facebook.
Users are still mad. But something interesting happened:
By Q1 2016 (just 2.5 years after introducing ads): Instagram generates $572 million in revenue in a single quarter. That’s 10% of Facebook’s entire revenue at the time.
By the end of 2016: $3.2 billion in total revenue for the year.
2024: Instagram generates over $66 billion in annual revenue. The platform has an estimated potential value of $200 billion. That’s 200 times what Facebook paid for it.
Current user count: Over 2 billion monthly active users.
⚠️ The Pattern
The users who threatened to leave stayed. The predicted mass exodus never actually happened. And Instagram today is just Instagram. With ads. And most people under 30 don’t even remember the controversy.
Reddit’s Anti-Corporate Identity
Reddit’s story hits different because being anti-corporate was core to their identity. The community took pride in this. Redditors would mock Digg for selling out. The ethos was: we’re different, we’re community-driven, we’ll never be like those other platforms.
November 2009: Reddit launches sponsored links.
The announcement tries to make it community-friendly: “Now for as little as $20, you can buy sponsored links on reddit: advertising by redditors, for redditors!”
The community’s reaction: hostile. Many users felt Reddit violated the social contract. Comment threads filled with accusations of selling out.
2010: Reddit launches Reddit Gold as a compromise. Premium subscription, ad-free experience, community features. The idea: give users a way to support the site without ads. It generates less than $1 million in revenue. Essentially a tip jar.
The site is bleeding money. Server costs are climbing. User base is growing. Revenue isn’t covering infrastructure for 200+ million monthly users.
2015: Reddit launches native ads (sponsored posts that look like regular Reddit posts). Revenue doubles.
Then watch what happens to revenue:
2018: $94 million
2019: $132 million
2020: $198 million
2021: $375 million
2022: $510 million
2023: $789 million
2024: $1.3 billion
Current stats: 97 million daily active users. The community is more engaged than ever. Ads account for over 90% of revenue. And nobody talks about Reddit selling out anymore. The “anti-corporate” platform runs on ads and nobody seems to care.
OpenAI’s Actual Options
OpenAI is burning around $9 billion in 2025, with projected losses of $14 billion in 2026. The company projects cumulative losses of over $100 billion before profitability. They won’t be profitable until 2029 at the earliest.
Given these numbers, they have three actual options:
Option 1: Destroy the free tier
Limit everyone to 5 messages per day. Use older, cheaper models. Make the free experience barely functional.
This drives users to competitors. Google Gemini grew 30% year-over-year in 2025. Claude grew 190%. Perplexity grew 370%. You lose market position. You lose the usage data that makes models better. You eventually lose everything.
Option 2: Keep burning
Maintain current quality and usage limits. Hope you can raise more money. Cross fingers that 2029 profitability actually happens. This leads to massive cumulative losses. Eventually investors stop showing up.
Option 3: Add ads
Add ads to free tier. Generate $1-3 billion in new annual revenue. Keep free tier quality high. Stay competitive.
For context on why this works: Spotify has 423 million users on ad-supported free tier. Generates $1.85 billion from ads annually. That’s only 11.8% of total revenue, but critically, 60% of premium subscribers started on the free tier.
More than a cost centre, they made free tier its top of the conversion funnel.
OpenAI picked option 3.
From OpenAI’s Announcement
- The model doesn’t know ads exist
- Sensitive conversations (health, politics, violence) get zero ads
- Conversations aren’t shared with advertisers
- Pro ($200/month) and Enterprise tiers see zero ads
- Their stated hierarchy: User Trust > User Value > Advertiser Value > Revenue
Could they break these promises later? Sure. But the framework is actually more restrictive than most ad platforms.
Claude’s Position Right Now
Claude can say “no ads” because they’re where Instagram was in 2012. 20-30 million monthly users. Serving developers and enterprises. 80% of revenue from API and enterprise customers, not consumer subscriptions.
They’ve raised over $37 billion total and are seeking another $20 billion. They’re burning cash too, just at a smaller scale with a different user mix.
They’re also deliberately avoiding expensive compute tasks. No video generation (which costs significantly more than text). Feature restrictions that keep costs manageable.
This works at 20-30 million users serving the 16% of early adopters. And if Claude ever scales to 300+ million users serving mainstream consumers (not just developers), they’ll face identical economics.
The VC funding won’t stretch forever. Enterprise revenue won’t cover consumer infrastructure at that scale. When Instagram hit 100+ million users, they needed ads. When Reddit hit 200+ million users, they needed ads.
If Claude hits those numbers serving mainstream users, they’ll need ads too.
Final Thoughts
AI compute scales linearly with usage. When you’re serving 900 million users who expect it free, the math solves itself. Platforms survive by solving unit economics, not by running better marketing campaigns about staying pure.
Give it three years, nobody will remember being upset.
