AI pricing isn’t a technical problem - it’s a human one
There’s a lot you end up thinking about as a startup founder.
Success, doom... and the question that connects those dots:
How do we actually make money - in a way that lets this become a self-sustaining business?
If you’re building in AI, that question is... not easy.
I've been thinking about this topic a lot for Ting. I've seen pricing through a lot of angles in my career. At adidas (perceived innovation, even just more logos on item), Yeezy and drops more for hype and scarcity, and at TikTok and Snapchat the strategy was to start with low-cost items to get a credit card on file then expand the range and value of goods over time. Not to mention inflate the algo on commerce posts to drive the behaviour. And finally in video games I learnt a lot about free-to-play models.
All of these models are still fair game in AI.
Funnily enough, I think even with AI, the greatest technology the world has ever seen, I predict the return of three pricing structures which are just so, inherently, human.
And that, kinda, makes sense…
I've written a more in-depth Substack on the topic that I'll link to at the bottom, but here's the TL;DR to avoid you needing to switch platforms.
A short summary of the thinking
AI models are converging faster than we expected. For most users in the West, three models dominate day-to-day usage - ChatGPT, Claude, and Gemini - and the gap between them is shrinking.
What matters just as much as their capabilities is where they live: Google and Amazon can afford to keep model costs low because AI strengthens their existing money-printing businesses. Pricing, in that context, becomes a strategic weapon, not a reflection of cost. You can see how OpenAI gets squeezed in the future without cloud, ads, search etc.
As intelligence becomes cheaper and more interchangeable, pricing stops being about intelligence itself. Instead, it shifts toward human factors: taste, trust, personality, and brand.
This isn’t new. Other industries have followed the same pattern. Cars share underlying platforms but diverge wildly in price based on identity and perception. A Fiat and a Ferrari do the same job, so does a Casio and Rolex; the difference is feel.
AI appears to be heading in the same direction, where perception and signalling drive value.
From there, three patterns emerge:
First, luxury AI becomes plausible. Not luxury because it’s smarter, but because of who uses it, what it signals, and what it costs. In economic terms, some AI products may behave like Veblen goods where the price itself is part of the appeal.
Second, fragmentation increases. Rather than one AI for everyone, we’ll see many AIs tuned for specific tribes, workflows, and values. Like Superhuman in email, these products won’t win by being objectively better, but by being deeply felt by a particular group.
Third, privacy becomes a class system. Society has already accepted “pay to remove ads” elsewhere. It’s likely AI follows the same path: one group pays for privacy, the other pays with data. That has Black Mirror implications, but it’s hard to ignore as a mainstream economic model.
Outcome-based pricing sounds attractive in theory, but it carries risks IMHO. When agents are paid per outcome, incentives shift toward optimizing metrics rather than user intent - the same dynamic causes ad-driven products to degrade over time.
All of this leads to a simple conclusion: as intelligence commoditises, pricing stops being a technical problem and becomes a human and more cultural one. And I kinda like that?
For products like Ting, that means familiar pricing models in the short term - subscriptions, tiers, usage limits - because predictability still matters to users and investors. But longer term, the real value will likely come from brand, community relevance, and trust in a world where intelligence itself is abundant.
Pricing won’t be solved by benchmarks (maybe until AGI anyway...).
It will be shaped by humans being human.
Full article here: https://chiefting.substack.com/p/how-do-you-make-money-in-ai
Any other smart brains got thoughts on this?
Welcome more opinion and critical feedback if you're still online shipping in the coming days!
Dan
Replies