Are tdoay's prices actually the ceiling? OpenAI added a $500 tier pretty recently while lowering capacity for their $200 tier. I wouldn't be surprised to see Anthropic doing something similar soon, particularly once they are a publicly traded company beholden to making shareholders happy.

The problem for me is that it's really hard to project what this industry is going to look like 6 months from now, let alone 6 years.

What's the endgame once the dust settles? Is it going to be a utility like water or electricity and the cheapest offer will prevail, in which case margins will be crushed? Or is Anthropic (or some other big AI lab) going to swallow everything and become the Google Search of LLMs, with everybody else fighting for scraps?

Or is it local compute, models becoming more efficient and hardware becoming cheaper (again) and companies hosting their own AI clusters like they would run SAP?

I think there is a plausible path where this valuation actually works out, but it's a narrow one IMO. I also want to acknowledge that I personally would prefer if it wasn't the case and a more distributed, open model prevailed.

The Chinese labs are not providing their open-weight models for free because they think it is good for humanity. They are doing it because they are severely chip constrained due to the Nvidia and ASML embargos and all available GPUs have to go toward training with very little left for inference [1]. So they are giving their models away for US and EU hosts to serve inference in order to still have a presence in the market.

It's not a great position to be in and it will still take a long time for the compute gap to close. In the meantime, Anthropic, OpenAI, and Google are essentially unchallenged.

[1] https://www.konstantinpilz.com/posts-publications/chinese-st...

The online crowd harping on about Chinese models is disconnected from reality imo.

I won't use a Chinese model. Why would i bother? I pay $100/m for the frontier model, understanding it's marginally better than much cheaper alternatives. I get so much value from the product it's not worth shopping around

Think of it like apple and android. Plenty of people pay 2k for their iPhone every few years when android at 1/4 the price is just as functional. Why? Because they get so much excess value from the device it's not worth shopping around. For most people it's their only personal computer. From booking airfare, to filling out their kids permission slips, it's a device they use 10+ hours a week. They don't bother, until a paradigm shift happens and something is 10x better than the iPhone. I imagine llms will follow the same trajectory

I think you’re right - it’ll be interesting to see how the behemoths navigate when new and cheaper memory manufacturing capacity comes online. Currently predicted to happen starting 2028. That is where most of the current hardware costs are concentrated.

On that front, I think DeepSeek is a dark horse. They are really odd - not chasing revenue, care more about unit profitability, are obsessed with driving down costs (like Elon back in the day). As the founder says, they are content with where they are, being the model that does 80% for a fraction of the cost.

I mention them because they are playing a different game and might be able to out-innovate everyone else.

Anthropic is not in the model business. They were smart and they are building productivity ecosystem business. Look at what claude artifacts are doing now -- inline multi-user editing, comments that users and/or models can respond and take action on, import/export from all common formats. It's at least a Google Docs killer if not an M365 killer.

And claude code as a harness and skill/plugin/mcp ecosystem is far more valuable than any one model.

And coward for long-running tasks, though they're slipping on that one because they're really trying hard not to make Claude an assistant. We'll see if they react to Muse and change course.

But the Chinese free-weight models? Not a competitor. If/when those companies (or others) build vertically integrated productivity stacks on those free models, then we'll talk. Microsoft certainly saw that opportunity but couldn't get out of their own way and are basically roadkill at this point.

Even outside of the Chinese models, Amazon, Google, Microsoft, Meta all of them are offering significantly more coherent and competitive internal models and reasonable price.

Why would I pick anthropic.

That depends on whether "frontier" models can maintain an expanding frontier. If the frontier settles, they're dead, and it will be commoditized. Otherwise, there are still plenty of pathological behaviors of these things to address that will keep expanding that frontier (if that's possible and the limitation isn't the approach itself).

The reception of saying things like this is interesting. AI stans think I'm way too pessimistic but AI haters think I'm mindlessly optimistic.

Cloud compute has managed to have pretty good margins. That seems like a really similar business model to what Anthropic is doing - inferene+value-adds as-a-service. Why wouldn't they also be able to get cloud level margins or a bit higher?

there is no point of saying 'the chinese' everyone will do it because it was proven that any political frills might cause an attempt to damage your system.

I dont know when, but i think if you keep menacing of taking technology of people hands for any political reason will be a reason for people to not use it or to rely in a more reasonable party.

I don't think today's prices are the ceiling, considering they are subsidized. Once businesses have grown dependent on AI (and their institutional knowledge is lost), they will pay much more just to stay afloat.

> Today's prices are the ceiling.

Why? I don't see any reason for that to be the case.

100% value fallacy right there.

Point is taken - competition is fierce and will be biting everyone's ankles.

Unfortunately it's now extremely likely that our governments will manipulate markets to keep frontier domestic AI afloat. They are financially and militarily dependent on domestic AI now.

Chinese models are being banned, will continue to be banned by USG. Even the Chinese govt is restricting external access to models now. It starts with banning Govt access to International AI, but will continue to consumer and business access.

buying anthropic = betting that there is going to be ai breakthrough at anthropic that will put its models ahead. AI stock market = polymarket

> Meanwhile the Chinese are going to undercut them every chance they get.

I wish they were, but I always wonder when people chime in with Chinese models what they are actually doing with those. I have an OpenRouter account and give various models a shot every other month, wishing for an alternative, but I always come back to Claude. Even more so since the switch from Opus 5.1 to 5.5. To me, personally, it's unmatched. Like a year ago it was very different, and I would use other models all the time. But honestly at the time they all felt kind of "meh" still. I just hope everyone is at a good baseline in a year from now on, so I actually have a choice. And then, as you stated, things will get interesting.

tl;dr: I agree