Discussion
Data centers were on the ballot and they lost big time. Voters in at least 29 jurisdictions approved restrictions on data centers.
Hyperscalers have committed trillions to AI infrastructure they may not be able to build.
The AI bubble is inflated by the idea that there’s enormous demand for AI. So what happens if the demand shows up but there’s no capacity to meet it? 🫧
@carnage4life the people “demanding” ai should build the data centers in their own back yards.
@carnage4life Time to put a stick of dynamite into big techs aspirations and reing them back in.
I do not understand what part of the economy of AI you are missing.
1) AI companies are losing money, big way;
2) building more data center will not lower the cost of compute, on the contrary;
3) for AI to be a viable, the cost of tokens must increase A LOT;
4) I do not see how raising the cost can increase demand, especially by looking at the current impact of AI on business productivity.
So, please show me a scenario in which demand increases.
@glipari It seems you don’t understand what the word demand means in this context.
All the major hyperscalers have already said there is more demand than they can handle TODAY including Amazon who said AWS is already “sold out” for 2027 and 2028 capacity is gone.
Maybe you need to do a bit more studying about the AI economy. 🤷🏾♂️
@carnage4life The major hyperscalers may have said so, but they can lie (or not, of course).
Whatever the hyperscalers say, when I need AI, I can still use it and for free. Therefore there must be free capacity left.
@carnage4life Is the demand even showing up?
@carnage4life probably bullish for the companies that have already built some decent compute
@carnage4life I mean given you are suggesting the demand hasn’t shown up yet perhaps this is the direct solution to “build it and they will come” and it is “don’t and they won’t”.