> which are known to be ridiculously expensive at 90% margins
Is that known?
Is that known?
Known to someone who doesn't trade on a heavily AI indexed market (in 2026 all companies, even Tyson or Chevron is an AI play but not as heavily AI indexed as those say 2 degrees away from Anthropic or OpenAI) - likely not.
So if the point you're making is your typical barber doesn't know about this, I am guilty.
If you want to do your own research into this, there are a lot of private reports and publications on the topic.
If you don't want to spend any money, there is even a recent public analysis using a 92% API gross margin.
Anthropic's October 5, 2026 analysis also points to a 92% API gross margin that SemiAnalysis based their report on.
OpenAI reported paid-user compute margin reached 70% in October 2025 - a year ago.
DeepSeek has similar reports. The frontier labs are comfortably in the fat inference margin territory.
Do you have information to the contrary?
It's interesting to read this data. A lot of the "common wisdom" was that subscriptions were heavily subsidized, but in actuality they're just not quite as outrageously profitable.
Both can be true! Assume that it costs you $10 to support inference of 1MM tokens, and you sell that for $100. So you have a 90% profit margin.
Now you also have flat rate subscriptions for $200/month, where you sell upto 100MM tokens. That costs you $1000, but you're selling it as a loss leader because that drives up demand, bottom up - for your product, which is good for long-term profitability specially when your investors are not yet sensitive to you burning cash.
I have multiple $500/mo OAI subs. Still cheaper than buying my own hardware for equivalent performance.