#AnthropicCEOCallsForAISlowdown
Model companies can’t slow down.
Anthropic alone has committed to expand its compute capacity to over 11.5 GW.
At the bottom end of the current rates, this means at least $115-$120 billion annual compute expense.
Assume similar numbers for OpenAI and we are looking at $1 trillion compute spending over the next 5 years.
How are they going to cover for this if they slow down?
It’s not possible.
Model companies can’t slow down.
Anthropic alone has committed to expand its compute capacity to over 11.5 GW.
At the bottom end of the current rates, this means at least $115-$120 billion annual compute expense.
Assume similar numbers for OpenAI and we are looking at $1 trillion compute spending over the next 5 years.
How are they going to cover for this if they slow down?
It’s not possible.
