OpenAI's Gabriel Petersson argues that prioritizing $20-a-month consumer AI over the $40 trillion enterprise market slows AGI progress
Story Overview
Gabriel Petersson, recently transitioned from OpenAI's Sora team, posted on X that consumer plans priced at $20 per month show little demand for frontier-level model performance, while enterprises collectively spend around $40 trillion yearly on knowledge work and actively seek those advances, making the consumer emphasis a potential drag on AGI development.
Scale mismatch shapes incentive structures
Petersson notes that matching enterprise-level spend through consumer subscriptions alone would require roughly 166 billion users, a figure that underscores why he sees enterprise priorities as better aligned with pushing model capabilities forward.
No quantified slowdown evidence appears
The post presents the consumer-versus-enterprise contrast as an illustrative opinion rather than new benchmark data or internal OpenAI metrics, leaving the actual pace impact on AGI efforts unmeasured in the available statements.
Many users endorse the push for enterprise and B2B focus in AGI development because it reveals what customers will actually pay for, while others react with sarcasm or doubt AI's capabilities.
Most Activity
Do you want to be liked (B2C) or do you want to win (B2B)
consumers pay 20$/month and don't care about frontier performance
enterprise pay $40T/year for intelligence (knowledge work), and really care about frontier performance
focusing on consumer is a mistake and anti-agi pilled
you need 166B people on 20$ subscriptions to get to the size of enterprise
and that ignores enterprise exploding in market size as ai get adopted
consumers pay 20$/month and don't care about frontier performance
enterprise pay $40T/year for intelligence (knowledge work), and really care about frontier performance
focusing on consumer is a mistake and anti-agi pilled
@gabriel1 does this mean as consumers we are going to lose fair and equal access to intelligence once frontier labs have enough enterprise customers?
@MaziyarPanahi if you think the iphone meant only rich people will ever have smartphones and it will never trickle down into 30$ phones then yes
@gabriel1 where does this 40T figure come from? convinced enterprise is big but that is like
@beffjezos Why not both? :)
@gabriel1 That's why more and more frontier labs are focusing on enterprice, for example now OpenAI :D
@xadisingh that's salaries for knowledge workers (source chatgpt 5.5 extended thinking)
@gabriel1 but i hope we agree the difference between the cheapest and the most expensive iphone is not on the same scale as the difference between ai models. with iphones, you can live with the cheapest one. but today, cheap ai models are pretty bad. that’s why we push for frontier models.
@gabriel1 And now we know which type of customer you're looking for in you new endeavor :) GL
@gabriel1 What about if you give a product so good to consumers that they pay more than $20? When you have custom hardware, it becomes a different dynamic more like what Apple has achieved.
@gabriel1 But I'm a consumer and I want SOTA...
@gabriel1 Google will take care of 20$/month thing, what we need from @OpenAI and @AnthropicAI is frontier performance, nothing else
@gabriel1 Source trust me bro?
@gabriel1 Sora Enterprise replacing Linkedin would have been killer!
@gabriel1 So you're saying they need to bump up birth rates?
@gabriel1 is “all of humanity” agi-pilled?
@gabriel1 I think glasses are the ultimate form factor for the universal assistant, and with this there are many gen ui opportunities which are on the frontier
other than that, i think the main benefits to agi come from enterprise (advancing science, von neumann machines)
@gabriel1 what's the margin on both?
@gabriel1 missing a few scale variables.