Connect your Claude, Codex or OpenClaw for full transcripts on millions of podcasts. Two week free trial, cancel anytime.
Start free trial →Unchained · Crypto & Web3 · November 2026
Kain Warwick on the gap between what AI labs charge on flat-rate subscription plans and per-token API prices for the same inference. He argues the labs are subsidising plans to grow usage, and that resellers will close the gap.
…the challenging part, right? If you could strip out the risky stuff using the blockchain, then you could just. Buy inference from whoever's selling the cheapest inference. And at the moment, you have this very, very crazy arbitrage that exists, which is the cost of API versus the cost of a plan.
Subscription is so much cheaper. Yeah. So
much cheaper, right? And you cannot have an extremely valuable asset that you're selling with price discrimination like this and not have the world show up and be like, whoa, whoa, whoa. Like we're going to collapse this delta, right? To the point where eventually, you know, the API and the plan has to, you can't subsidize something this much and not have people turn up and be like, well, we're going to arm this subsidy, right? And so the idea is that like there's a bunch of people for a bunch of reasons that are willing to sell you inference cheaper than what the labs are willing to sell you inference at their API rack rates, right? And they will distribute it via the blockchain. And so even
if you're breaking terms of service or anything, whatever, man. They're still going to do it. Yeah.
It's not even, they're just making a two-sided marketplace, right? You got supply over here and you've got demand over here. And we're just going to let supply and demand have a party in the middle. And like, I mean, it's, it, you know, we're seeing this with like Venice. I think Venice has done a…
But why?