> And, while most data centers need to be located near to their key users, to reduce latency, AI data centers are often located where energy and land are cheap
I think it's often overstated how important it is for data centers to be near users, including in this piece. Just look at the actual latency numbers, e.g., https://wondernetwork.com/pings/San%20Francisco. From SF to the East Coast is sub 100ms (round-trip), from SF to New Orleans is ~51ms. This is just not enough to matter for *most* LLM uses we see today; real time things like translation or voice are the exception. Maybe uses will change, but I expect the vast majority of LLM use won't be latency sensitive on the order of 1/20th of a second.
Some really good points in here. But I am puzzled at why so many people including the author are prepared to dismiss the “$160B/320B per year over a decade” possibility.
If you add up all direct AI ARR (openAI/anthropic/cursor/etc) that number will have gone from below $10B to greater than $30B over the course of 2025. (This is before even adding things like Meta revenues, given that some of their massive advertising income is at least partially provided by GPU workloads. Can set that aside for the moment even though this might make total AI revenues many multiples higher).
It would not surprise me if these direct AI revenues went from $30B to $75B to $150B to $250B to $375B or something over the next five years, given the way the AI models are improving and spreading via increasingly strong business applications. Why then would we view the prospect of $1.6T-3.2T over a decade as outlandish? If anything it strikes me as quite likely that we clear that hurdle.
> And, while most data centers need to be located near to their key users, to reduce latency, AI data centers are often located where energy and land are cheap
I think it's often overstated how important it is for data centers to be near users, including in this piece. Just look at the actual latency numbers, e.g., https://wondernetwork.com/pings/San%20Francisco. From SF to the East Coast is sub 100ms (round-trip), from SF to New Orleans is ~51ms. This is just not enough to matter for *most* LLM uses we see today; real time things like translation or voice are the exception. Maybe uses will change, but I expect the vast majority of LLM use won't be latency sensitive on the order of 1/20th of a second.
For reference on how much 50ms is, ChatGPT as measured by Artificial Analysis (https://artificialanalysis.ai/models/gpt-5-chatgpt/providers) has ~530ms time to first token (latency), many small models are also in the 0.3-1s range.
Some really good points in here. But I am puzzled at why so many people including the author are prepared to dismiss the “$160B/320B per year over a decade” possibility.
If you add up all direct AI ARR (openAI/anthropic/cursor/etc) that number will have gone from below $10B to greater than $30B over the course of 2025. (This is before even adding things like Meta revenues, given that some of their massive advertising income is at least partially provided by GPU workloads. Can set that aside for the moment even though this might make total AI revenues many multiples higher).
It would not surprise me if these direct AI revenues went from $30B to $75B to $150B to $250B to $375B or something over the next five years, given the way the AI models are improving and spreading via increasingly strong business applications. Why then would we view the prospect of $1.6T-3.2T over a decade as outlandish? If anything it strikes me as quite likely that we clear that hurdle.
Perfect timing, this. The hype cycle is unsustainabile.