This probably isn’t a very accurate article. But I still find it plausible anyway. I feel like most AI subscriptions have only now started showing value to me. And even then it’s because I’ve been willing to work and experiment with it to get it to do things on my behalf.
The vast majority of users are likely comfortable with just asking it quick questions, and probably don’t need agents running in the background all the time.
That's kinda the point: The vast majority have no intense use for LLMs that would justify them paying a monthly subscription. But if they don't subscribe, it's not sustainable!
A somewhat misleading article or at least headline as this is about households which will not account for people paying for AI subscriptions at work unless they are using their own personal card.
Imagine creating something people still like to pay for after over 20 years. Incredible! Disclaimer: I also have still a subscription.
Regarding LLM: I guess if you count in enterprise/company subscription, LLMs would win. Interesting would be how many really want them or just have to use them…
According to this [1] data which uses actual subscriptions instead of households, it seems the number is around 8-10million LLM subscribers in USA. This was March so I expect the figures to be around 40% higher now.
Another proxy: OpenAI said [2] they have 50M subscribers and in another place they said 40% of mobile revenue comes from USA. That itself accounts for 10M paid subscribers within USA which is 5 times the estimation from this post using ChatGPT alone, that too from February.
Open AI is fudging its wording by saying more than "9 million paying business users rely on ChatGPT for work, and startups, enterprises, and governments" -- IE they're blending together paying consumers with businesses. Also not setting a time frame, so no idea how many of them are recurring users.
Other source is questionable since (for starters) it relies on sampling and mentions a 10% inaccuracy (!) rate.
The 2-3% number is from two banks, based on direct payment card data.
These are numbers that even skeptics are not skeptical of. Worth toning down your skepticism. It’s a bit nutty to dismiss a neutral poll as well as disclosed numbers by the companies - two different sources for two different companies. At some point it must start dawning on you
OK well Zitron is citing The Information which is generally solid, but you're kinda burying the lede. According to this, it's 9 million subscribers because...
"OpenAI projects an 80% decline... from 44 million in 2025 to 9 million in 2026" [!!!]
Surely there's reason to be skeptical when OpenAI's own internal projection indicates a massive, massive loss in subscribers?
I'm still inclined to believe OpenAI's projection was too optimistic. The bank report says that average subscriptions only last 7 months, so that would also help explain the rapid churn.
BTW, the number of global WoW subscribers is currently about... 9 million.
> OpenAI projects an 80% decline... from 44 million in 2025 to 9 million in 2026
I'm curious to know why you think OpenAI is predicting that their subscribers would decline. Why would the company be so bearish about their own financials?
This probably isn’t a very accurate article. But I still find it plausible anyway. I feel like most AI subscriptions have only now started showing value to me. And even then it’s because I’ve been willing to work and experiment with it to get it to do things on my behalf.
The vast majority of users are likely comfortable with just asking it quick questions, and probably don’t need agents running in the background all the time.
That's kinda the point: The vast majority have no intense use for LLMs that would justify them paying a monthly subscription. But if they don't subscribe, it's not sustainable!
I guess that WoW is more fun to play than using LLMs. Guess what will always have more costumers than LLMs?
> more costumers
Well it seems easier to dress up as a Night Elf than dress up as ChatGPT. At least, that’s my impression looking around San Diego Comic-Con footage.
Eh, combo of mobile keyboard autocorrect and non native speaker but yes, comics are fun.
A somewhat misleading article or at least headline as this is about households which will not account for people paying for AI subscriptions at work unless they are using their own personal card.
True, but personal use tracks direct consumer adoption, whereas sustained enterprise use is still very much in the air. (See "tokenmaxxing".)
1) This could not have been farthest from truth
2) WoW will never be billed via a corporate card.
Imagine creating something people still like to pay for after over 20 years. Incredible! Disclaimer: I also have still a subscription.
Regarding LLM: I guess if you count in enterprise/company subscription, LLMs would win. Interesting would be how many really want them or just have to use them…
This is false or misleading.
According to this [1] data which uses actual subscriptions instead of households, it seems the number is around 8-10million LLM subscribers in USA. This was March so I expect the figures to be around 40% higher now.
Another proxy: OpenAI said [2] they have 50M subscribers and in another place they said 40% of mobile revenue comes from USA. That itself accounts for 10M paid subscribers within USA which is 5 times the estimation from this post using ChatGPT alone, that too from February.
[1] https://epoch.ai/data/polling/2026-03
[2] https://openai.com/index/scaling-ai-for-everyone
Open AI is fudging its wording by saying more than "9 million paying business users rely on ChatGPT for work, and startups, enterprises, and governments" -- IE they're blending together paying consumers with businesses. Also not setting a time frame, so no idea how many of them are recurring users.
Other source is questionable since (for starters) it relies on sampling and mentions a 10% inaccuracy (!) rate.
The 2-3% number is from two banks, based on direct payment card data.
These are numbers that even skeptics are not skeptical of. Worth toning down your skepticism. It’s a bit nutty to dismiss a neutral poll as well as disclosed numbers by the companies - two different sources for two different companies. At some point it must start dawning on you
https://www.wheresyoured.at/openai-projects-chatgpt-plus-sub...
OK well Zitron is citing The Information which is generally solid, but you're kinda burying the lede. According to this, it's 9 million subscribers because...
"OpenAI projects an 80% decline... from 44 million in 2025 to 9 million in 2026" [!!!]
Surely there's reason to be skeptical when OpenAI's own internal projection indicates a massive, massive loss in subscribers?
I'm still inclined to believe OpenAI's projection was too optimistic. The bank report says that average subscriptions only last 7 months, so that would also help explain the rapid churn.
BTW, the number of global WoW subscribers is currently about... 9 million.
> OpenAI projects an 80% decline... from 44 million in 2025 to 9 million in 2026
I'm curious to know why you think OpenAI is predicting that their subscribers would decline. Why would the company be so bearish about their own financials?
You cannot fudge your books. At the end of the day we will see if the AI cult is caught swimming naked.
I was there when pets.com happened.