RESEARCH PUBLICATION – Artificial Intelligence – AI Ecosystem part II – OpenAI

28th November 2025: RFM updates its coverage of Artificial Intelligence with the publication of Artificial Intelligence – AI Ecosystem part II – OpenAI. RFM Subscribers will receive their copy by email.

The AI ecosystem is a must-win for OpenAI. Apple and Google have demonstrated that owning a digital ecosystem can be extremely lucrative and result in a multi-trillion-dollar valuation. OpenAI needs all of this if it is to make good on its promises. OpenAI has the users, but it must now execute its developer program to ensure that it becomes the go-to place to use, buy or sell an AI service.

• Open AI’s big promises: With $1.4tn being spent on infrastructure over the next 5 years, OpenAI needs to generate more than $500bn in annual revenues to come close to earning a positive return for its stakeholders. This is why it must win the consumer and enterprise AI ecosystem race, as without these victories, it will badly miss this target.
• LLM as a platform: LLMs are becoming commodities, which means that there is little point in developers building them from scratch. Instead, RFM expects OpenAI, Google, Anthropic, Meta and so on to offer their LLMs as platforms upon which developers can build. This is the race for the AI ecosystem, and there will be many casualties.
• The AI Ecosystem: is the digital environment where users (buy side) spend their time and where sellers (sell side) of AI services can create, deploy and sell their services. When the environment becomes the go-to place to transact, there is a very large and secure monetisation opportunity for the owner of the ecosystem, which in this case is OpenAI.
• Flywheel effect: which is where a growing user base attracts more developers, which in turn attracts more users and so on. The flywheel is notoriously difficult to get moving, but this must be achieved by any successful ecosystem.  
• Buy Side: where OpenAI is already way ahead of its rivals. RFM estimates that ChatGPT already has around 1.2bn MaU, more than three times its nearest credible rival, Gemini. OpenAI is already well past the tipping point at which the user count becomes a major incentive for developers to create AI services for ChatGPT.
• Sell Side: is more of a challenge. OpenAI has claimed that it has 4m developers, but with its tools only just launched, this figure is not very credible, and RFM considers its developer program to be nascent. OpenAI has no experience with developers, while its chief rival Google has been working with developers for over 20 years. OpenAI cannot get this wrong.  
• Open AI Assessment: OpenAI fares reasonably well on RFM’s qualitative assessment. In terms of its user experience, it does well but with everyone using a text prompt, this is not hard to copy. The two areas where it is weak are in 3^rd party developers where it is very early days, and in its economic viability. How OpenAI will make money with a $1.4tn investment plan remains a mystery, and if the money dries up it will be in real trouble.