Here is something most people are not saying directly: the AI tools you have been paying $20 to $100 per month for do not actually cost that little to run.
OpenAI loses approximately $1.22 for every dollar it earns. Anthropic filed its IPO at a $965 billion valuation while burning cash at scale. Both companies have been subsidizing your access to their models with investor money. Venture capital money. Money that will, at some point, need a return.
Both companies filed for public markets within days of each other. OpenAI filed a confidential S-1 with the SEC on June 8. Anthropic filed on June 1. OpenAI targets a September 2026 debut. Anthropic is not far behind.
This is not doom-and-gloom. But it is a change in conditions, and leaders who think clearly about their tools act accordingly.
What Actually Changes When These Companies Go Public
Public markets reward profit. That is not a cynical observation. It is the structure of how public capital works.
Right now, both OpenAI and Anthropic answer primarily to large private investors who have patience for a long game. Once public, they will answer to quarterly earnings expectations. Analysts. Shareholders who did not sign up to lose $1.22 per dollar at scale indefinitely.
The pressure to close the gap between revenue and cost will accelerate. The most direct path to doing that is pricing.
Not necessarily through dramatic, immediate price increases. More likely through the gradual removal of subsidies. The elimination of “included at no extra cost” features that quietly get moved to premium tiers. The reduction of free plan generosity. The restructuring of enterprise contracts at renewal time.
We are already seeing signals. Claude Fable 5, released today, is free on current plans through June 22. That is a 13-day window. After June 22, Anthropic has not confirmed it will remain in current tiers. That is the future of AI pricing — generous windows followed by pricing normalization.
The Era We Have Been Living In
It is worth naming something clearly.
Every business that has built AI into its operations over the past two to three years has been benefiting from a historically unusual subsidy. The tools we have been using were priced not to make money, but to build market share. Every dollar we spent on AI was matched — and likely exceeded several times over — by venture capital filling in the gap.
We built our workflows, our automation systems, our content engines, and our client deliverables on top of pricing that was never designed to reflect the true cost of what we were using.
That is not a criticism. It was a rational decision given the conditions. But the conditions are changing.
The question now is: what did you build that has real value regardless of what AI tools cost, and what did you build that only works because AI was cheap?
What the Entrepreneurs Who Will Thrive Are Doing
The shift to public markets is not a crisis for most small businesses. It is a clarifying moment.
Here is how to think about it:
Audit your AI spend against outcomes. Go line by line. For every AI tool you subscribe to, identify the specific business outcome it produces. Revenue generated, time saved with a dollar value attached, client deliverables improved. If you cannot name the outcome, you are paying for potential, not performance. Cut those first.
Build workflows, not dependencies. A workflow is yours. It produces a specific output, and you understand how to get that output. A dependency is a tool you use without knowing what it actually does in your process. Dependencies become liabilities when prices change. Workflows are portable and adaptable.
Lock in what you can. Annual contracts are priced before price increases. If you are month-to-month on tools you know you will use for the next 12 months, switching to annual now locks in current pricing before IPO pressure hits. This is not about panicking. It is about making a rational decision with information most people are not acting on yet.
Document your prompts. The prompts that produce your best AI outputs are intellectual assets. If the model changes, the platform changes, or the pricing changes and you need to move, documented prompts transfer. Muscle memory does not.
Do not wait for a competitor to force the issue. The entrepreneurs who will navigate this shift best are the ones who make intentional choices now, before conditions require it. Not because they predicted the future perfectly, but because they paid attention and acted.
The Deeper Strategic Principle
There is a broader leadership pattern here that extends past AI pricing.
Whenever you build your operations on a resource that someone else is subsidizing, you are exposed to their risk. This is true of AI pricing. It has been true historically of cloud pricing, social media organic reach, payment processing rates, and supplier relationships.
The answer is not to refuse to use subsidized resources. The answer is to use them aggressively while they are cheap, build real capabilities on top of them, and maintain enough flexibility to adapt when the subsidy ends.
The entrepreneurs who got crushed when Facebook organic reach declined in 2014-2016 were the ones who had built entire marketing strategies on free distribution without building an owned audience in parallel. The AI equivalent is building automation workflows without building the human knowledge and judgment that the automation is supposed to amplify.
AI tools going public does not mean they stop being useful. It means the cost of accessing them will, over time, reflect the actual cost of running them. Build your business on the outcomes AI enables, not on the assumption that AI will always be cheap.
What I Am Watching
A few specific things I am tracking as both of these IPOs move through the process:
Enterprise contract structures. Once public, how both companies restructure enterprise contracts at renewal will be the clearest signal of where consumer pricing goes. Enterprise leads. Consumer follows.
Free tier changes. Both companies have generous free tiers that lose money. The rate at which those tiers are restructured or eliminated is a leading indicator of pricing philosophy post-IPO.
The pricing gap between OpenAI and Anthropic. Both companies are racing to public markets. If one raises prices faster, the other gains market share. This competitive dynamic could actually keep prices lower than they would otherwise go. Worth watching.
New entrants. Google, Meta, Mistral, xAI, and open-source models all benefit when OpenAI and Anthropic are forced to prioritize profitability over market share. Commoditization is real. If a good enough open-source model runs locally at near-zero marginal cost, pricing pressure on the incumbents gets complicated quickly.
Frequently Asked Questions
Will AI prices increase dramatically right after the IPOs?
Probably not dramatically right away. Public companies typically manage analyst expectations carefully in their first few quarters. The more likely pattern is gradual reduction of subsidized features and moderate price increases over 12-24 months, not a sudden jump.
Should I switch to cheaper AI alternatives now to hedge?
Only if a cheaper alternative actually meets your needs. Switching tools has real costs in workflow disruption and re-learning. The better hedge is documenting your current workflows well enough to make switching feasible if needed, rather than switching preemptively.
Does this affect the open-source AI model ecosystem?
The open-source ecosystem benefits when proprietary models raise prices. Models like Llama and Mistral get more competitive as the cost gap narrows. Local models have improved dramatically. For entrepreneurs who can run local models, this is worth revisiting in the next 6-12 months.
Is this a good or bad sign for AI overall?
Both companies going public is a sign of maturity and institutional confidence in the AI sector. It is a normal stage in a technology’s development cycle. It is not a bad sign. It is a signal to be thoughtful about how you are building on top of the technology.
What is the difference between OpenAI’s and Anthropic’s financial situations?
OpenAI reports $20B+ annualized revenue with a loss ratio of approximately $1.22 per $1 earned. Anthropic filed at a $965B valuation backed by $35B in chip financing. Both are burning cash significantly. OpenAI’s losses are larger in absolute terms given higher revenue. Anthropic has a higher valuation relative to its stage.
Should small businesses be concerned about AI pricing?
Concerned is not the right word. Attentive is. If AI tools are core to your business operations, understanding the pricing trajectory and building flexibility into your workflows is good risk management. It is not an emergency. It is prudent planning.
Published by Jonathan Mast | jonathanmast.com
Sources: SEC filings, Anthropic investor materials, CNBC, Financial Times — June 2026
