If you've been using ChatGPT to boost your income – whether that's writing freelance articles, building apps, creating content, or helping run your side hustle – OpenAI has a message for you: they might want a piece of the pie.
In a blog post published this week, OpenAI's Chief Financial Officer Sarah Friar laid out the company's evolving business strategy, and buried in the corporate language was a bombshell that has the tech world buzzing. As artificial intelligence moves deeper into "scientific research, drug discovery, energy systems, and financial modeling," Friar wrote, "new economic models will emerge. Licensing, IP-based agreements, and outcome-based pricing will share in the value created."
What OpenAI Actually Said
To be fair, let's look at exactly what Friar wrote in her official company statement. The CFO explained that OpenAI's business model is designed to "scale with the value intelligence delivers." So far, that's meant consumer subscriptions ($20/month for ChatGPT Plus), workplace subscriptions, and API pricing based on usage.
But here's where things get interesting. Friar explicitly stated that the company's future revenue models will go "beyond what we already sell." She pointed to industries where AI is making increasingly significant contributions — drug discovery, scientific research, energy systems, financial modeling — and said that "licensing, IP-based agreements, and outcome-based pricing will share in the value created."
The phrasing is deliberate. "Share in the value created" is corporate-speak for taking a percentage of the profits that users generate with AI assistance.
The Drug Discovery Example That Has Everyone Talking
Think about this scenario:
Under an "outcome-based pricing" model? OpenAI could potentially negotiate a percentage of the drug's revenue or a substantial licensing fee tied to the commercial success of the AI-assisted discovery.
The same logic applies to other high-value use cases. A novelist who uses ChatGPT to help outline, edit, and polish their manuscript into a bestseller. A startup founder who leverages AI to build a product that gets acquired for millions. A financial analyst who uses AI-powered modeling to make trades that generate substantial returns.
In each case, OpenAI's current business model captures almost none of that value. A flat $20 or $200 monthly subscription doesn't scale with success. But a revenue share arrangement? That's a completely different equation.
Why This Makes Business Sense for OpenAI
Let's be real: OpenAI is burning through cash at an almost incomprehensible rate. According to recent reports, the company spent over $8 billion on inference costs alone in 2025 – the expense of actually running the AI models that power ChatGPT. Despite bringing in over $20 billion in annual revenue (a figure Friar confirmed in her blog post), OpenAI posted a net loss of $13.5 billion in just the first half of 2025.
OpenAI is losing roughly three dollars for every dollar it earns.
And here's the uncomfortable reality that subscription pricing can't solve: only about 5% of ChatGPT's 800 million users actually pay for the service. The company is essentially subsidizing free usage for 95% of its user base with revenue from the paying minority.
Meanwhile, some of those free users are generating significant value with the tool. A freelance writer earning thousands of dollars per month from AI-assisted content. A small business owner using ChatGPT to handle customer service, create marketing materials, and streamline operations. A developer building and selling apps with AI-generated code.
From OpenAI's perspective, the current model leaves money on the table. Lots of it.
What "Outcome-Based Pricing" Could Actually Look Like
While OpenAI hasn't announced specific implementation details, we can make educated guesses about how revenue-sharing models might work based on precedents in other industries.
For enterprise and pharmaceutical applications, this probably means negotiated licensing agreements. A biotech company using AI for drug discovery might sign a deal that includes a percentage of milestone payments or royalties if an AI-assisted compound makes it to market. These arrangements are common in the pharmaceutical industry – it's just that now AI companies want to be at the table.
For smaller-scale commercial users, the model could look more like what we see in e-commerce platforms or app stores. Think of how Shopify takes a transaction fee on sales, or how Apple and Google take 30% of app revenue. OpenAI could implement a similar system for commercial applications built on its API, taking a percentage of revenue generated by AI-powered products.
For individual users, the implementation gets trickier. Would OpenAI try to track whether you used ChatGPT to write a book that became profitable? Monitor whether your freelance income increased after you started using the tool? The technical and ethical challenges here are substantial, which is probably why Friar's comments focused on enterprise use cases like drug discovery rather than individual creators.
But don't assume individuals are off the hook. The language about "sharing in the value created" is broad enough to encompass any commercial use of the technology. And as AI becomes more capable of generating entire products – not just assisting with them – the line between "tool" and "co-creator" becomes increasingly blurry.
The Backlash Has Already Begun
News of OpenAI's revenue-sharing ambitions spread quickly through tech communities, and the reaction has been... mixed, to put it mildly.
Critics point out the obvious tension:
Users are already paying for the service. If you're shelling out $20 or $200 per month for ChatGPT access, shouldn't that cover your right to use the tool commercially? Adding a success tax on top of subscription fees feels like double-dipping.
Others raise concerns about fairness and attribution. How would OpenAI determine what portion of a success was attributable to AI assistance versus human creativity and effort? If a researcher uses ChatGPT to help organize their thoughts but does the actual scientific work themselves, does OpenAI deserve a cut of any resulting discoveries? Where do you draw the line?
There's also the competitive angle. Google's Gemini, Anthropic's Claude, Meta's open-source Llama models, and dozens of other AI tools are all competing for users. If OpenAI implements aggressive revenue-sharing requirements, users who are generating significant commercial value might simply switch to alternatives with more favorable terms.
And let's not forget the legal questions. Copyright law is already struggling to keep up with AI-generated content. Adding revenue-sharing agreements to the mix creates a whole new layer of complexity. Who owns AI-assisted work? What rights does the AI company have to profits generated from that work? These questions haven't been definitively answered by courts yet.
The Freelancer's Dilemma
For the millions of freelancers, content creators, and small business owners who have integrated ChatGPT into their workflows, this news raises uncomfortable questions.
Consider a freelance writer who uses ChatGPT to speed up research, outline articles, and polish drafts. They're earning, say, $5,000 per month from their writing work. Under a revenue-sharing model, would they owe OpenAI a percentage of that income?
The practical challenges are enormous. How would OpenAI even know what you're earning? Would users be required to self-report commercial income? What happens if you refuse? Could OpenAI audit your usage and retroactively claim a share of past earnings?
These questions don't have answers yet, which is probably intentional on OpenAI's part. Friar's comments signal the direction the company is thinking, but the implementation details remain fuzzy. This gives OpenAI flexibility while also creating uncertainty for users trying to plan their businesses around AI tools.
For now, the safest assumption is that any revenue-sharing arrangements will start with large enterprise customers and high-value commercial use cases. Individual freelancers probably aren't going to get a bill from OpenAI for 10% of their earnings anytime soon.
But "probably" and "anytime soon" are doing a lot of heavy lifting in that sentence.
How This Compares to Other Tech Business Models
OpenAI isn't the first tech company to explore outcome-based pricing. In fact, the model has precedent in several industries.
Consulting firms have long used success fees, where they take a percentage of cost savings or revenue improvements they help clients achieve. Investment banks charge fees based on the size of deals they help close. Even real estate agents work on commission tied to the value of transactions.
In the AI space specifically, companies like IBM have experimented with outcome-based pricing for enterprise AI services. The idea is that if AI delivers measurable business value, the vendor should share in that value – it aligns incentives and makes the technology more accessible to companies hesitant to pay large upfront fees.
What makes OpenAI's situation different is scale. IBM might negotiate individual enterprise contracts with outcome-based pricing. OpenAI has 800 million users and counting. Applying any kind of revenue-sharing model at that scale presents unprecedented logistical and ethical challenges.
The Broader Implications for the AI Industry
If OpenAI moves forward with outcome-based pricing, it could reshape how the entire AI industry approaches monetization.
Right now, most AI companies follow a straightforward model: charge for access, whether through subscriptions or API usage fees. This treats AI as a tool, like a word processor or spreadsheet application. You pay for the software, and whatever you create with it is yours.
Outcome-based pricing treats AI differently as something closer to a collaborator or partner. If the AI meaningfully contributes to a valuable output, it (or rather, its owner) deserves a share of the proceeds. This isn't just a business model change; it's a philosophical shift in how we think about AI's role in creative and commercial work.
The implications extend beyond business models to questions of ownership and credit. If OpenAI takes a percentage of profits from AI-assisted drug discoveries, are they implicitly claiming partial credit for those discoveries? What does that mean for scientific attribution? For patent applications? For Nobel Prizes?
These questions might sound far-fetched today, but AI capabilities are advancing rapidly. Last year's hypothetical is this year's practical consideration.
What Users Should Do Now
If you're using ChatGPT for commercial purposes, here's my practical advice:
First, don't panic. OpenAI's comments are about future business models, not immediate changes. Your current subscription terms remain in effect, and any significant policy changes would require notice and likely face substantial pushback.
Second, read the terms of service carefully. OpenAI's current terms already include provisions about commercial use, though they don't include revenue sharing. Stay informed about any updates or changes to these terms.
Third, diversify your AI toolkit. Don't become completely dependent on any single AI provider. Explore alternatives like Claude, Gemini, or open-source models. Having options gives you leverage and reduces risk if any one provider implements unfavorable terms.
Fourth, keep records of your creative process. Document what work you do yourself versus what AI assists with. This could become important if ownership or attribution questions arise in the future.
Fifth, follow the news. AI policy is evolving rapidly, both from companies and governments. Staying informed helps you adapt as the landscape changes.
The Bottom Line
OpenAI's CFO has put the industry on notice: the company is actively exploring ways to capture more of the value that ChatGPT helps users create. Whether that means revenue sharing on commercial products, licensing fees for enterprise discoveries, or something else entirely remains to be seen.
For now, this is more of a signal than a policy. But it's a signal worth paying attention to.
The era of flat-rate AI subscriptions may not last forever. As these tools become more capable and more central to how we work, create, and earn – the companies behind them will naturally seek business models that better capture the value being generated.
The question isn't whether AI companies will pursue outcome-based pricing. It's when, how, and whether users will accept it.
For millions of people who have come to depend on ChatGPT for their livelihoods, that question suddenly feels very personal.
Frequently Asked Questions
Is OpenAI currently charging users a percentage of their earnings from ChatGPT?
No. As of now, OpenAI uses subscription-based pricing ($20/month for ChatGPT Plus, $200/month for ChatGPT Pro) and usage-based API pricing. There is no current revenue-sharing arrangement with individual users. The CFO's comments describe potential future business models, not current policies.
What exactly did OpenAI's CFO say about revenue sharing?
Sarah Friar wrote that as AI moves into areas like "scientific research, drug discovery, energy systems, and financial modeling," new business models will emerge, including "licensing, IP-based agreements, and outcome-based pricing" that will "share in the value created." She did not announce specific implementation details or timelines.
Would freelancers and content creators have to pay OpenAI a percentage of their income?
There's no announcement about this specifically. Friar's comments focused on enterprise-level applications like drug discovery. However, the language about "sharing in the value created" is broad enough to potentially encompass any commercial use. Implementation for individual users would face significant practical and legal challenges.
How would OpenAI even know how much money I make using ChatGPT?
This is one of the major unanswered questions. For API users, OpenAI could potentially track usage patterns and require disclosure of commercial applications. For individual ChatGPT users, enforcement would be much more difficult. Any revenue-sharing program would likely start with large enterprise customers where tracking is more feasible.
Can I switch to a different AI if OpenAI implements revenue sharing?
Yes. Alternatives include Anthropic's Claude, Google's Gemini, Meta's Llama (open-source), and many others. Competition in the AI market gives users options if any single provider's terms become unfavorable. However, switching may require adjusting your workflows and learning new tools.
Does this mean OpenAI would own part of my creative work?
Not necessarily. Revenue sharing is different from ownership. OpenAI isn't claiming copyright or ownership over AI-assisted work — they would be charging for the commercial value that their tool helped create. However, the legal distinctions between contribution, ownership, and licensing in AI-assisted work are still being defined.
When would these changes take effect?
No timeline has been announced. Friar's blog post described strategic direction, not immediate policy changes. Any significant changes to pricing or terms would require notice to users and would likely be tested with enterprise customers before broader implementation.
How much of a percentage could OpenAI take?
This is entirely speculative since no specific figures have been announced. For context, app stores typically take 15-30% of revenue, while consulting success fees can range from 10-40% depending on the engagement. Enterprise AI licensing deals vary widely based on negotiated terms and the specific value being delivered.
Will this affect the free version of ChatGPT?
The free version of ChatGPT is already supported by advertising (recently introduced) and effectively subsidized by paying users. Revenue-sharing arrangements would most likely apply to commercial or professional use rather than casual free-tier usage. However, OpenAI's terms could evolve as its business model develops.
What should businesses using ChatGPT API do to prepare?
Businesses should review their current API agreements, document their AI usage and the value it creates, stay informed about terms of service updates, and consider building flexibility into their systems to potentially switch providers if needed. For high-value applications, consulting with legal counsel about AI-related intellectual property issues may be prudent.
Related Articles
Mark
Dani
Dani
