Log in
Venture capitalists Menlo Ventures have released what purports to be a survey: “2025: The State of Consumer AI”. That is, chatbots. [Menlo Ventures]
The subtitle is: “AI’s Consumer Tipping Point Has Arrived”! That’s a very positive statement! But it’s not one these numbers actually show.
Menlo engaged Morning Consult to survey 5,031 Americans. Here are the takeaways:
- There are actually a ton of ordinary US consumers who have used a chatbot at least once in the past six months! 20% use it daily!
- This is because the chatbots are free and convenient.
- Consumers use chatbots for casual things. They’re toys.
- Only 3% of chatbot users pay anything. Though that number’s a bit weird.
Let’s take the numbers seriously — probably more seriously than Menlo did. Let’s assume the sample is a fair sample, and this actually reflects how the American populace uses chatbots.
More than half of American adults (61%) have used AI in the past six months, and nearly one in five rely on it every day.
That’s a big number! Why do ordinary people use chatbots? They’re convenient. They’re just there. Google has chosen to just not work as a search engine any more. Gemini is free, Google is shoving it in your face.
There’s a few odd details. The people they asked can’t tell the large language model chatbots like ChatGPT or Gemini from the old keyword bot assistants like Alexa and Siri. They’re calling all of these “AI.”
But dig this logical leap:
This is no longer experimentation; it’s habit formation at an unprecedented scale.
Maybe? Or maybe it’s a toy and they use it because it’s free.
Then Menlo takes these US numbers and multiplies them by the whole population of the Earth. As if everyone else on the planet is a variety of American:
Scaled globally, that translates to 1.7–1.8 billion people who have used AI tools, with 500–600 million engaging daily.
That number — not the US totals that Menlo measured — is the number they proceed to speculate on wildly:
1.8 billion users at an average monthly subscription cost of $20 per month equals $432 billion a year; today’s $12 billion market indicates that only about 3% pay for premium services — a strikingly low conversion rate and one of the largest and fastest-emerging monetization gaps in recent consumer tech history.
So even the “3%” number has problems. Menlo did not ask in the survey if people paid for their chatbot out of their own pocket. They just surmised a number after the fact. If this is Menlo’s big talking point number, why didn’t they even ask if survey respondents were subscribers?
If you’re wondering about that writing style, “Claude Sonnet 4” is credited as one of the authors.
That phrase “largest and fastest-emerging monetization gap” is what Menlo’s blog post is about. That’s its takeaway. Menlo puts all its effort into making out that AI’s massive failure to sell to ordinary people is just a market opportunity — and not evidence that the market of paying customers doesn’t want to buy yet another monthly subscription for a casual toy.
Menlo puts Generative AI at a $12 billion paying market so far — but compare that to the investment into generative AI, on the order of half a trillion dollars so far just set on fire.
For comparison, Ed Zitron figures the total stated revenue for big tech generative AI at $8 billion if you don’t count what OpenAI spends at Azure, with a stated $327 billion spent by the biggest techs. So Menlo’s figures seem to be of the right magnitude. [Ed Zitron]
Menlo wants to push that there are market opportunities here for their portfolio companies! How can they get this alleged 97% of casual chatbot users to cough up yet another monthly fee?
Well, maybe they can’t. Generative AI is a toy. And it’s only free because venture capital firms like Menlo subsidise it so heavily.
What happens when the venture subsidy stops flowing and the chatbots suddenly have to pay their way? What happens when the prices multiply 5× to 10× and there’s no free tier? Are you going to casually muck about with a chatbot at $200 a month? Probably not.
Menlo has an answer — they look forward to turning chatbots into a sea of spam:
We expect rapid adoption of advertising models, transaction fees, affiliate revenue, and marketplace models.
If you have to cajole employees into using the AI, then you need to consider that maybe chatbots don’t work well enough that people would use them casually if they had to pay — that the “97%” free users are not in fact likely to start paying out their own pockets. And you just set all that venture money on fire. And it’s not coming back.
