Q: Apparently the price for computation by DeepSeek R1 is $0.55 per million tokens compared to $15 per million tokens for OpenAI o1, which it is comparable to in terms of performance. This is about 97% cheaper.
I’ve been reading about Jevon’s paradox ie. as something gets cheaper, usage explodes.
But to my math usage will have to go up 33x to cover the difference in revenue to Nebius.
This will eventually happen, but short-term, aren’t we looking at revenues falling off a cliff for Nebius?
I’ve been reading about Jevon’s paradox ie. as something gets cheaper, usage explodes.
But to my math usage will have to go up 33x to cover the difference in revenue to Nebius.
This will eventually happen, but short-term, aren’t we looking at revenues falling off a cliff for Nebius?