Ep. 036 - $200 Buys $12,000 of Opus Tokens, We Bought Every Plan (AI Cloud TCO) | Max Kan, Jordan Nanos, Andrew Megalaa

Max Kan (0:05)

Hello, everyone.

Jordan Nanos (0:06)

Welcome back to SemiAnalysis Weekly. This week, I'm joined by Max and Andrew, the authors of the very controversial, but very interesting Anthropic subscriptions offer five times or more value than OpenAI. This was an article that we put out a couple of days ago, where the guys did limit testing for every AI subscription plan. It's not just OpenAI with Codex and Anthropic with Claude Code. It also includes Meta, SpaceX AI, MiniMax, Moonshot, ZAI, Cursor, Cognition, anybody where you could be getting an all-in plan from.

And we're going to talk about the learnings that they've had. Guys, welcome to the show.

Max Kan (0:45)

Thank you for having us, Jordan.

Jordan Nanos (0:48)

Okay, so first-

Max Kan (0:49)

I feel like this is a TV interview or something. It's like the setup is too college now. I'm not used to it.

Jordan Nanos (0:57)

Well, yeah, okay. Other people need me to direct them with questions, but for you, I think we can let you run it. Max, the plans are highly subsidized.

Max Kan (1:06)

No.

Okay, it depends on what you actually- depends on how you want to define subsidization. Relative to API pricing, the plans are obviously a better deal. I mean, we can throw up some graphics on the screen here, or maybe people can read the news letter if they haven't, but some headline results, including you being able to pay like $200 for a cloud plan and get like $12,000, over 55 tokens at API pricing. So that's obviously a good deal.

However, in terms of whether or not the subscriptions are like subsidized in an absolute sense, like do you get- does Anthropic have negative gross margins when they're selling a subscription plan? The answer is no. The reason for this is that A, their margins at API pricing are incredible. I mean, consensus is like over 90 percent blended at this point. Then the second thing is that most people don't use like 100 percent utilization. So if you combine like a low sort of realistic average utilization, where I think it's around like say 10 percent, with this 90 percent plus margin at API pricing, you end up with subscriptions having like, call it 50 to 60 percent like blended gross margins for Anthropic. So it's obviously still actually a pretty good business for them. But compared to their like 90 percent plus margin API business, that's search are still so.

Jordan Nanos (2:34)

So maybe the takeaway is not that the subscription plans are subsidized, but that the API pricing for the current frontier models is like punitive, exploitative, price gouging, whatever word you want to.

Andrew Megalaa (2:48)

It's quite expensive.

Max Kan (2:50)

They're selling to willing buyers at the fair market price. This is capitalism.

Jordan Nanos (2:59)

Understood. Okay. So talk to me about the 100 percent utilization and how you actually did the testing to realize these numbers, these results.

Max Kan (3:11)

That's an Andrew question. Yeah, Andrew.

Andrew Megalaa (3:15)

Yeah. So I guess to figure out the results for the subscription plans, we're pretty much just like we buy every single plan and ended up measuring every single token type.

From that, we're able to say what kind of workload we are running, in this case, generally agentic. That's pretty much why I'm using these subscription plans for. They'll tend to be around 96-ish percent cash reads, a little bit of inputs and a little bit of cash try around 2 something percent, and a little bit of output. You blend that together and then you can get the big numbers that we figured out.

Max Kan (3:53)

Sorry, before you dive more into the details, I do want to emphasize that there have been people on tour before who did similar analysis, and they would essentially just present one dollar value as the equivalent API value of subscription. That is just a fundamentally oversimplified and correct analysis.

It is the way to think about your subscription plan is that by giving the company $200 a month or whatever, they allocate you some number of credits. This is just some other currency. And you should think of each model token type combo as costing a different number of credits, and essentially drawing from your allotted usage of a different amount. And because the ratio between different model token type credit costs can be different from their API price ratio, the actual API equivalent value of your subscription plan can change dramatically depending on what model you're using and what workload you're running. And what this means is that you can't just make a blanket statement like, oh, the $200 Anthropic plan gives you $10,000 of value. You have to say the $200 Anthropic plan, when running a particular workload on a particular model, gives you some dollar amount. And this is the nuance that we capture in our newsletter, and we want to make sure that people understand.