Why Sam Altman is Wrong About ChatGPT’s New PRO Plan Pricing
Why Sam Altman is Wrong About ChatGPT’s New PRO Plan Pricing
This week, Sam Altman, OpenAI’s CEO and founder, publicly revealed that OpenAI is losing money on its $200/month PRO plan, hinting that users are “using it too much.” (Bloomberg article)
While having users “use a product too much” is a good problem to have—and I suspect Altman’s declaration was meant to highlight the value of the PRO plan—the idea that ChatGPT PRO is underpriced and costing OpenAI money doesn’t hold up. The narrative that “on average, PRO users are not paying enough compared to what they use” is flawed, and here’s why the pricing strategy deserves a rethink:
Key Arguments:
Contribution Margin is Likely Positive (AI Costs Are Mostly Fixed, Not Variable)
- The major expense in AI comes from training the models, not their usage. This means costs are largely fixed.
- With higher usage, it might be tempting to say, “We need to increase prices to recover costs.” However, markets are not completely inelastic. Plus, “a lot of usage” doesn’t necessarily translate to “a lot of costs.”
- If the goal is to recover costs, the logical approach might be to reduce prices, increase volume, and better leverage the fixed-cost structure—not raise prices.
Averages Are Misleading in Pricing
- Altman’s suggestion that “on average” PRO users are overusing the service is based on flawed reasoning. Averages are poor indicators for pricing strategy.
- Take car buyers: the average spend on a new car might be $20K. But in reality, you have Skoda buyers at $10K and Tesla buyers at $40K. None of them is an “average” user.
- Pricing must account for user segments and the actual value delivered, not arbitrary thresholds or averages.
The Price Gap Between $20 and $200 is Huge
- There’s an enormous gap between the $20/month Plus plan and the $200/month PRO plan. It’s true that many users are getting more than $20 of value from ChatGPT—but between $20 and $200, there’s a world of difference!
- Heavy users are not necessarily the ones with the deepest pockets or those deriving the most value. The correlation isn’t straightforward.
- Many users are stranded between these two tiers, unwilling to commit to such a significant jump in price.
Flat Increases Won’t Work in a Competitive Market
- Even if raising prices could work in theory, the competitive landscape today makes it impractical.
- Microsoft, Anthropic, and other players now offer viable alternatives to ChatGPT. The era of “price skimming” for AI services is over.
- Competitors providing similar or better services will quickly capture users unwilling to pay a premium based solely on OpenAI’s reputation or hype.
But I believe Altman the problem is elsewhere and Altman himself gave us a strong hint of the issue there.
Altman admitted that the pricing decisions for Plus and PRO weren’t based on rigorous studies but on personal judgment: “I basically just thought they were good.”
Don’t get me wrong—Altman & Co. are brilliant when it comes to AI. But pricing isn’t just about intuition; it requires structure, segmentation, and data-driven decision-making.
The bigger picture and the future of AI Pricing
The days of pricing based on AI hype and early-stage scarcity are coming to an end. Starting in the second half of this year, the “charge more because it’s AI” bubble will burst.
Companies, including OpenAI, will have to adopt more reasonable pricing strategies, focusing on the actual value they deliver rather than the allure of cutting-edge technology.
Scary for those who like setting prices by tossing a coin, isn’t it?
Takeaways:
- Averages Are Not Reliable Metrics: Over-reliance on averages to guide pricing decisions leads to flawed conclusions.
- Huge Price Gaps Alienate Users: The gap between $20 and $200 excludes many potential subscribers.
- The AI Price Skimming Era is Over: Competition is here, and pricing based on hype alone won’t be sustainable.
- Value-Centric Pricing is Key: Future pricing will need to align with the actual value created for users—not just the allure of AI.
- Whoever you are and doesn’t matter how good your product is – Don’t just toss a coin to set your prices!
My name is Salva, I am a product exec and Senior Partner at Reasonable Product, a boutique Product Advisory Firm.
I write about product pricing, e-commerce/marketplaces, subscription models, and modern product organizations. I mainly engage and work in tech products, including SaaS, Marketplaces, and IoT (Hardware + Software).
My superpower is to move between ambiguity (as in creativity, innovation, opportunity, and ‘thinking out of the box’) and structure (as in ‘getting things done’ and getting real impact).
I am firmly convinced that you can help others only if you have lived the same challenges: I have been lucky enough to practice product leadership in companies of different sizes and with different product maturity. Doing product right is hard: I felt the pain myself and developed my methods to get to efficient product teams that produce meaningful work.