It's a misleading number, the 500 billion in revenue is needed over a 10 year period, it's not a yearly figure. Anthropic isn't paying 500 billion a year for the commitments that OP posted, it's paying 500 billion between now and 2036.
FWIW I max out my pro max plan basically every week, not even using Fable any more. I think it's just a matter of people figuring out how to apply tokens effectively.
I don't think they are betting on that many people subscribing to their $200/mo plan, there's pretty much no upside in using a frontier model for the average consumer.
But on the flip side I don't think aiming for 1M businesses spending $40k a month in token in average is going to be more viable. And I'm not sure having 1000 corporations spending $40M a month is a option either.
This is the issue, for two reasons: Is there trillions of dollars of demand for LLMs, which is what it takes for Anthropic to make $500 billion in revenue? Is that demand going to be met by frontier cloud LLMs, or by open weight models run by the organizations using them (and cheaper cloud alternatives)?
It's a misleading number, the 500 billion in revenue is needed over a 10 year period, it's not a yearly figure. Anthropic isn't paying 500 billion a year for the commitments that OP posted, it's paying 500 billion between now and 2036.
[dead]
In the simplest calculation, if you get 200M people on a $200/mo pro max plan, that's $500B annual revenue.
This is a "utility" mindset. If you bet that Anthropic is just like a monthly electric bill, then it's within the realm of possibility.
For the record - I don't think that's gonna happen.
FWIW I max out my pro max plan basically every week, not even using Fable any more. I think it's just a matter of people figuring out how to apply tokens effectively.
I don't think they are betting on that many people subscribing to their $200/mo plan, there's pretty much no upside in using a frontier model for the average consumer.
But on the flip side I don't think aiming for 1M businesses spending $40k a month in token in average is going to be more viable. And I'm not sure having 1000 corporations spending $40M a month is a option either.
How many thousand more datacenters would that need, and what would be the ongoing (massive) cost of deprecating hardware in that world?
This is the issue, for two reasons: Is there trillions of dollars of demand for LLMs, which is what it takes for Anthropic to make $500 billion in revenue? Is that demand going to be met by frontier cloud LLMs, or by open weight models run by the organizations using them (and cheaper cloud alternatives)?
[dead]