FWIW, I asked Claude Opus 4.5 in research mode to attempt to do this per-model-ROI analysis for OpenAI, and then for Anthropic, from what public materials it could locate, and it seemed to think that even in this framework OpenAI’s ROI is deeply negative: primarily because a) training run investment includes not only the final successful run but also failed runs (the same issue as in the numbers DeepSeek released) b) revenue earnings are depressed by competition so are not much above serving costs, and c) model depreciation cycles are viciously short, generally less than 6 months.
So, even on a per-model ROI basis, OpenAI are still in a “burning VC money to gain market share and intellectual capital” mode.
Of Anthropic, it seemed to think their per-model ROI was also still negative, but less so for a variety of reasons (fewer failed training runs, slower model obscolescence), and was improving. It found their predictions of profitability by 2028 plausible. (I didn’t ask it whether it might be biased.)
However, in an AI slowdown, factor c) automatically improves, and there are fairly obvious levers OpenAI could pull to improve a) and b) — some of which apparently Anthropic are already pulling.
For both companies, it mentioned that users in their highest individual subscription tiers often have usage so high that they lose them money. So I expect we’ll eventually see tighter usage caps and even higher subscription tiers.
Why would factor c automatically improve if half a year is more than enough for the Chinese companies to catch up by distilling frontier models and release their open-weight models which then will be served almost at the cost of inference? If anything, in this scenario the models themselves will be fully commodified and margins in AI will be determined by the product characteristics
I guess I’m implicitly assuming any AI slowdown is for technical reasons, so also affects Chinese companies. If it were, say, a loss of investor confidence in LLM foundation labs, then it might not affect Chinese companies if they had sufficient CCCP state funding.
If it’s for technical reasons, then it should hit Chinese companies only as soon as they catch up, doesn’t it? I’m not sure I understand your argument.
Also, I don’t think Chinese companies have any viable business model for future scaling anyway since no one outside China wants to send their data on Chinese servers. Hence they are forced to economize as much as possible, and it’s possible they are supported by Chinese authorities for political reasons.
I take your point. I think I was assuming as slowdown, not a wall, and that the Western companies had more money so got further. But yes, it’s quite sensitive to circumstantial details.
FWIW, I asked Claude Opus 4.5 in research mode to attempt to do this per-model-ROI analysis for OpenAI, and then for Anthropic, from what public materials it could locate, and it seemed to think that even in this framework OpenAI’s ROI is deeply negative: primarily because a) training run investment includes not only the final successful run but also failed runs (the same issue as in the numbers DeepSeek released) b) revenue earnings are depressed by competition so are not much above serving costs, and c) model depreciation cycles are viciously short, generally less than 6 months.
So, even on a per-model ROI basis, OpenAI are still in a “burning VC money to gain market share and intellectual capital” mode.
Of Anthropic, it seemed to think their per-model ROI was also still negative, but less so for a variety of reasons (fewer failed training runs, slower model obscolescence), and was improving. It found their predictions of profitability by 2028 plausible. (I didn’t ask it whether it might be biased.)
However, in an AI slowdown, factor c) automatically improves, and there are fairly obvious levers OpenAI could pull to improve a) and b) — some of which apparently Anthropic are already pulling.
For both companies, it mentioned that users in their highest individual subscription tiers often have usage so high that they lose them money. So I expect we’ll eventually see tighter usage caps and even higher subscription tiers.
Why would factor c automatically improve if half a year is more than enough for the Chinese companies to catch up by distilling frontier models and release their open-weight models which then will be served almost at the cost of inference? If anything, in this scenario the models themselves will be fully commodified and margins in AI will be determined by the product characteristics
I guess I’m implicitly assuming any AI slowdown is for technical reasons, so also affects Chinese companies. If it were, say, a loss of investor confidence in LLM foundation labs, then it might not affect Chinese companies if they had sufficient CCCP state funding.
If it’s for technical reasons, then it should hit Chinese companies only as soon as they catch up, doesn’t it? I’m not sure I understand your argument.
Also, I don’t think Chinese companies have any viable business model for future scaling anyway since no one outside China wants to send their data on Chinese servers. Hence they are forced to economize as much as possible, and it’s possible they are supported by Chinese authorities for political reasons.
I take your point. I think I was assuming as slowdown, not a wall, and that the Western companies had more money so got further. But yes, it’s quite sensitive to circumstantial details.