AI Ecosystem – Breaking Barriers

Google opens a new front against OpenAI and others.

• Google is leveraging its experience in Android to remove as many barriers as possible to users switching from other AI agents to Gemini in an early move to take the one thing it does not already have from OpenAI: users.
• Google has launched two new tools that aim to make it as easy as possible to switch to Gemini by making it easy to import personal data and preferences, as well as the entire chat history, into Gemini.
• For those who use generative AI for inquiries, drawing pictures and research, the chat history is rapidly becoming a library of useful information which is unique to each AI service provider.
• One can quickly see how this would become a barrier to switching in the same way as WhatsApp or iMessage history is a barrier to switching from Android to iOS or vice versa.
• Google has long made tools available to make it as easy as possible to port one’s data from iOS to Android, and the release of these tools is precisely the same strategy aimed at attracting users to Gemini and away from everyone else.
• OpenAI is still by far the leader when it comes to the AI ecosystem, but Google is increasingly snapping at its heels.  
• Furthermore, Google has certain advantages, such as massive cash flow from its existing businesses and its ownership of the Android platform, that it can use to push its AI services to users.
• OpenAI has the advantage of being the first mover in this field, being a household name, and, on my subjective measures, it tends to perform better on some research tasks.
• However, its greatest weakness is that if the market decides that it has had enough of pouring billions of dollars into the company in the hope that there is a pot of gold at the end of the rainbow, then the company will cease to exist.
• If this happens to Google, then its share price will go down, but it will still be generating cash and in a position to continue investing to win the AI ecosystem.
• At the moment, it is very easy to export one’s chat history from ChatGPT into a pdf file, and I suspect that we will quickly see two reactions from OpenAI (and everyone else).
• First, I expect to see OpenAI quickly launch switching tools of its own and second, I expect that OpenAI will make it harder to export one’s data from ChatGPT so that it can be easily incorporated into Gemini or anyone else
• This is just the latest front in the developing battle to win the consumer AI ecosystem, and trillions of dollars are at stake.
• Other fronts include enticing developers to develop AI services for users on their in-house LLMs as well as offering free tiers of service.
• Google’s entire search business will one day move in this direction, and OpenAI’s entire business plan rests on its ability to earn money from consumers by serving up answers to their enquiries.
• This is why OpenAI is so focused on advertising, as 800m+ of its users consume substantial quantities of compute capacity but do not pay it a dime for the pleasure.
• To be successful at IPO, the company needs to show a clear and credible pathway to earning hundreds of billions of dollars in revenue from users, and advertising is a credible way of doing that.
• However, I am not convinced that it will be enough, as Google managed to earn around $400bn in 2025, whereas OpenAI will need $600bn or more just to cover the depreciation charge on the compute capacity that it is building.
• This is why I remain very cautious with the prospect of paying $1tn+ for OpenAI at IPO unless it suddenly invents a super intelligent machine that renders the human race obsolete.
• There is no sign of this, and so I remain very comfortable with my cautious stance on the company.
• I would choose Google over OpenAI any day of the week.