AI Is Now about Infrastructure
March 16, 2026
Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.
I listened to an interview conducted by Dwarkesh Patel with his “roommate” Dylan Patel. My question is, “Can roomies have an objective, unbiased, and fact-based interview?”
Thanks, Midjourney. Good enough with the roomies’ thing.
The premise of “Dylan Patel — Deep Dive on the 3 Big Bottlenecks to Scaling AI Compute. Plus, Why an H100 Is Worth More Today Than 3 Years Ago” is that AI is no longer a software problem. It is a power and infrastructure problem. I agree, however, I am not sure that I can accept some of the unsupported assertions presented in the interview. I suppose roomies don’t have to ask one another tough questions. At the end of the interview, the two have to return to the same home, share the facilities, and avoid the type of friction that created what I think of the Rob Reiner problem.
Please, listen to the one hour interview or read the transcript of the program. I want to offer one quote that struck me as illustrative of the approach:
If you look at what Anthropic has done over the last few months, with $4 billion or $6 billion in revenue added, we can just draw a straight line and say they’ll add another $6 billion of revenue a month. People would argue that’s bearish, and that they should go faster. What that implies is they’re going to add $60 billion of revenue across the next ten months. At the current gross margins Anthropic had, as last reported by media, that would imply they have roughly $40 billion of compute spend for that inference, for that $60 billion of revenue.
Dylan Patel, one roomie, is the “leadership” of SemiAnalysis consulting firm. His approach to smart software is what I call a kinder, gentler Ed Zitron approach. Both want to be noticed. Both present some interesting generalizations. Both thrive when the social media world notices their observations.
What are the main points of the Dylan Patel showcase?
I jotted down four:
First, Anthropic is going to be a big AI winner. My response is, “Maybe?” But $60 billion in revenue when Anthropic reported $6 billion to the US government. Yeah, maybe?
Second, the US economy is going to grow faster. My response is, “But the war may have a somewhat negative impact?” Geo-politics does not come into play in this interview when the US economy is discussed.
Third, the statements about China are speculative. The phrase, like the reference to a “roomie,” shot in the dark undermines the credibility of China-related statements.
Fourth, the assertion that ASML can make machines that produce the chips needed for AI is interesting. ASML does not make a toaster. Its “machines” are complex systems. These depend on supply chains that produce components that are not on Amazon’s warehouse shelves. With the Iranian war, assertions about ASML delivering machines and seamless production of chips are disconnected from reality.
Net net: Big talking and wild statements are part of what makes the name of Dylan’s SemiAnalysis appropriate. “Semi” thinking and partial analysis. Roommate interviews provide insight into curating enthusiasm, not probing assertions. Is that why these fellows are roomies?
Stephen E Arnold, March 16, 2026
Comments
612 Responses to “AI Is Now about Infrastructure”
Got something to say?
-
- Subscribe to Beyond Search
Feature archive
News archive
-
????? ?? ????? ?? ???? — ??? ?????? ??????????? ?????? ??? ??????? ??????? ?????????? ? ????????? ??? ???????? ?????????? ? ?????????????? ??????????? ??????? ? ?????????? ????????. ????? ?????? ????????? ?????? ???????????? ??? ???????? ? ??????? ???????? ?? ???????? ????????? ? ????????????????. ? ??????????????? ??????? «??????? ????? 24» ????? ??????????? ???????????? ????????????? ? ??????? ??????????? ? ?????? ???????? ????????? ????????. ????????? – ????? ?? ????? ?? ???? ???? ?????-?????????
? ???? ?????? ?? ??????? ??????? ?????????????? ????? ????????????, ?????????? ???????? ?? ????????? ??????? ? ???????? ? ????????????. ???????? ??????, ??? ??????? ???? ????????????? ? ???????????? ???????? ??????? ??? ?????????? ?????????? ?????????.
??? ??????? ??????????? – ?????????? ?? ???? ????????????
?????????? ?? ???????? ? ????????? ????? ? ?????? ???????????? ????? ??????????? ??????????????? ????????? ??????????? ??????????? ???????? ???????????? ???????? ?? ???????? ????. ???? ??????????? ???? ??????? ????? ????? ?????????? ????? ?????????? ?????????? ??? ??????? ????????. ?????? ?????????? ??????? ?? ????????? ???????? ????????? ???????? ? ??? ?????????????? ????????????. ??? ????????? ?????????? ??????????? ?????????? ? ??????????? ??????? ??????? ???????? ????????? ?????? ??????????????. ????????? – ?????????? ?? ???????? ?? ???? ? ??????
??? ?????? ???????? ?????????????? ? ????????????? ???????, ??? ?????? ?? ????????? ??? ????????? ???????? ????????????. ?? ?????????? ????????????? ??????? ????????? ??? ? ??????????? ??? ????? ??????, ??????? ????? ????????? ????????? ? ???????.
??????? ??? ?????? – ????? ????????? ?? ???
?????????? ?????????? ??? ??????????? ???????????? — ?? ??????? ???? ?? ??????????? ?????????? ?????. ?? ???????, ??? ????????? ??????? ???????????? ? ?????????, ? ?????? ????? — ? ????? ??? ????????????? ?????????.
? ???? ?? ???????????? – ?????????? ?? ???? ?? ???????? ??????? ????
?????????? ???????? ????????? ? ?????????? ????????? ? ????????????. ?? ???????????? ??????? ?? ??????????? ???????? — ?? ???????, ??? ??????? ??????????? ???? ? ????????????? ? ?????????? ??????????? ?? ???? ????.
?????? ?????????? ????? – ?????????? ?? ????????
???????? ?? ???????? ?????? ??????? ????????????? ???????? ? ????????????? ??????????? ??????????. ??? ???????????? ????????????? ???????? ? ???????????? ????????. ???????????? ? ???????? – ????? ????? ????????? ?? ???
??????? «??????? ??????» ? ?????????? ?????????? ??????????? ???????????? ????????? ??????? ???????? ??????????? ??????? ???????????? ???????? ?????? ??????????? ? ????????? ?????????????? ?????????. ??????? ???????? ??????????? ?????????????? ????? ????????? ? ?????? ???????? ????? ???? ????? ???????????? ???????? ? ??????????? ???????????. ????????? ?????????? ? ???????? ?????? ?????????????????? ? ??? ?????????? ??????????? ????????????. ??????????? ??????? ???????????? ???????????? ???????? ??????? ????????? ??????? ?????? ? ?????????? ????? ?? ?????????? ????? ??????? ??????????????. ????????? – ????? ?? ????? ?????????? ?? ???? ?????????
? ?????? ????????? ????? ????????? ?? ??? ???????????? ??? ?????????? ????????? ??????????? ??????????? ?????? ? ?????? ???????. ???? ????????? ? ???????????? ??????????? ???????? ?????? ? ????????? ???? ??????? ?? ?????. ????? ?????? ????????? ????????? ????? ?? ?????? ?????? ? ??????? ???????? ?? ????????. ????????? ??? – http://narkologicheskaya-pomoshh-nizhnij-novgorod-8.ru/
? ???? ?????? ?? ????????????? ?????????????? ??????? ??????????? ?? ????? ????????. ??????????? ???????, ????? ??? ????????, ????????? ? ???????????????? ??????????. ???????? ?????? ? ????????????? ????????? ?? ??????? ? ? ????? ? ??????????????.
? ??? ??????? – ?????????? ? ????????
????? ????? ???????? ??????? ? ???????????? ???????? ???????????? ?? ????????? ? ??????????? ?????????????????. ? ???????? ??????? ??????????? ??????? ???????? ?????????????? ???????? ?? ?????????? ???????? ? ????????? ?????????????? ????? ??? ??????????????? ??????? ??????????? ?? ?????? ??????????? ??????????? ? ?? ???????????? ???????? ????????. ????? ?????? ?????????? ????? ??? ?????????????? ?????????? ??????????? ??????????? ??????????? ? ???????????? ?????????? ???? ??? ?????????? ????????????????? ??????. ???????? ?????????????? ???????? – ????? ?? ????? ? ?????????? ? ?????-??????????
????? ?? ????? ?? ???? ? ?????????????: ?????? ??? ????? ?????? ???????????? ? ??????????? ????????? ? ??????????????? ??????? «??????» ????????? ????? ?????? ??? – ????? ?? ????? ?? ???? ???????? ? ?????????????