Hacker News new | ask | show | jobs
by qaq 37 days ago
If people were not consuming their services they would not be buying inference hardware at this rate so it's pretty much on consumers.
3 comments

They are reserving future HW productions to meet their hypothetical usage as well. Which is why others (like Apple) can’t reserve it for their future products.

Yet the AI labs are speculating on usage, and spending money from investments without clear revenue path.

Yes 65B ARR that Anthropic has is clear indication there is no path to revenue.
Sorry, I should have said "profit path", good catch! They have revenue, but their cost scales with revenue and they're losing more than they are making.

See: https://www.wheresyoured.at/brokenomics/ for an interesting write-up on the economics of AI.

Their costs do not scale linearly with revenue. Inference is expensive, but it's a variable cost. Anthropic's overall costs include massive fixed costs in training, which are the same regardless of usage.

It's easy to falsify the claim with a simple experiment: imagine they had no customer at all, $0 in revenue. Their costs would still be massive. If the claim were true, $0 revenue should mean $0 costs, right?

If people are sure they can always short NVIDIA
How much money does that revenue cost though? If I had to steel-man GPs argument I'd ask for profits rather than revenues.
We will see once they go public Dario did claim profit margin on inference is 40% if memory serves me right
That's convenient accounting. The reality is that they can't stop training since they risk losing customers if they do so. So they shouldn't factor it out of profitability analysis.
A lot of factors there we will see how it plays out.
Yes Dario is well known for his honesty
hence the bit about us learning the actual state of things once they are a public company.
This is not sustainable forever unless their hypothetical usage is realized, and eventually the bill will come due.

Meanwhile, component makers will surely be spinning up more capacity, some of them in a foolhardy manner, and if the bubble does burst, 3-6 months later we'll be seeing fire sales on components and component makers going bankrupt (or getting bailouts, if considered of national importance)

I feel like the fact Apple raised their prices means they foresee this lasting a lot longer than 3-6 months.
This is going to be the 1st increase of a series of increases. I don’t think this will ease in the next 2-3 years.
People will consume a lot of things offered below actual cost thanks to VC and cheap loans.

Doesn't mean people would legitimately use them enough to warrant such infrastracture demand, if they were priced according to actual costs.

So it's a distorted market.

Most of Anthropic revenue looks to be companies paying for Claude Code at API prices ...
Companies will consume a lot of things offered below actual cost thanks to VC and cheap loans.
API pricing is def not below cost
"def" doing a lot of work here.

It is more expensive? yeah

Is it "definitely not below cost"? hardly

except if as cost here only the inference cost is considered, and not the capital investment, and maintenance costs (not to mention r&d, marketing, and others).

to put it another way, if they just had the corporate API subs today, would they be profitable?

Ask every Windows 11 or Google consumer that doesn't give a damn for AI and, yet, has been almost forced to use Copilot and Gemini…