AI is turning into the worst possible business setup for AI startups. A commodity that requires huge capital investment and ongoing innovation to stay relevant. There’s no room for someone to run a small but profitable gold mine or couple of oil wells on the side. The only path to survival is investing crazy sums just to stay relevant and keep up. Meanwhile customers have virtually zero brand loyalty so if you slip behind just a bit folks will swap API endpoints and leave you in the dust. It’s a terrible setup business wise.
There’s also no real moat with all the major models converging to be “good enough” for nearly all use cases. Far beyond a typical race to the bottom.
Those like Google with other products will just add AI features and everyone else trying to make AI their product will just get completely crushed financially.
For consumers, the chat history is the moat. Why switch to a different provider for a marginal model improvement when ChatGPT already “knows” you? The value of sticking to a single provider is already there, even with the limited memory features they’ve implemented thus far.
I'm a heavy user of ChatGPT, and this is exactly why I haven't switched. I frequently search my old chats, or pick up one I started weeks or even months ago.
There is clearly a very strong moat. OpenAI is close to 1 billion active users on ChatGPT while Claude barely have any non-business users. Even though Anthropic had better models at different times this year, I never stopped using ChatGPT and paying for Plus.
We just don't know who will win in which area yet. It doesn't mean there is no moat.
I don’t think it’s a question of moat. The usage limits on the chat interface with the more advanced Claud models are brutal. I feel like I can barely start a conversation before I get shutdown. However, I switched over to Gemini almost completely and barely ever checkin with ChatGPT these days.
I can't prove it (because they show no stats) but I feel like you get more Gemini Pro with a normal subscription than Opus with a Max plan.
Maybe the new more efficient models made it better for Claude users but that was my experience a couple months ago.
For professional usage though, Calude Code is so much ahead of Antigravity that it didn't even make sense to make a formal comparison. That, even when using the same model (Opus).
OpenAI has close to 1 billion users which are mostly free users and will switch provider the moment OpenAI start charging them or adding ads. Which they will, as OpenAI themselves said they are losing money even with 200$ subs. So that amount of users is pretty meaningless.
Google and Microsoft have immense money printing machines. They can lose many billions of dollars for years and be fine as a business. OpenAI, not so much.
All of these have ads. And none of these have an equal value alternative. OpenAI, Claude, Deepseek, Mistral, Gemini, are mostly the same to a regular user.
Bing is mostly the same. Kagi is mostly the same. Yahoo, Yandex, etc. It's 2025. Hardly any difference. There were tens of search engines in the 90s and 2000s that were generally mostly the same. Yet, Google still won and owned nearly all search monetization.
Search was even easier to switch. At least ChatGPT has memory.
Most chat apps are the same as Whatsapp. All of them are free too.
"Ask ChaGPT" is the equivalent to "google it" in 2025.
No those are all significantly worse products, or at least were for a long time. I don't think OpenAI has anything close to a moat. They don't even have a short fence.
If you think of it like cloud, where it's a commodity that reaches competitive prices, then you can use it to build products and applications, instead of competing for infrastructure (see also: railroads, optical fiber)
There is tons of money to be made at the application layer, and VCs will start looking at that once the infrastructure layer collapses.
Not really though. The cloud has some stickiness. It’s pretty hard to move once you’ve settled in. For a lot of AI integrations though it’s just swapping some API endpoints and maybe tweaking the prompting a bit. For probably 95% of AI use cases there almost no barrier to switching.
Well, Claude has the best personality in a field where the rest are in a race to make the most awful personality. That's kind of a moat. The models were smarter too though the others have largely caught up, especially Gemini.
There’s also no real moat with all the major models converging to be “good enough” for nearly all use cases. Far beyond a typical race to the bottom.
Those like Google with other products will just add AI features and everyone else trying to make AI their product will just get completely crushed financially.