So what I mainly want to do is just post a string of like 52 confetti emojis, because this was definitely the best single piece of reporting on data centers I've seen anywhere. But I'm going to attempt to push past that urge and highlight the most actionable part of all of this, which I think (as you point out) what the entire dynamic reveals about *thymos* — and specifically, Silicon Valley's complete absence of both skill and interest in cultivating it. I don't think that phenomenon is by any means exclusive to SF; in many ways, we're simply entering the third decade of the 'Unwinding' Packer talks about as we've sorted and polarized the macrosocial/macroeconomic forces that shape America out and away from most of America. The Bay is simply one such force.
But what I think causes data centers to be such a irresistible lightning rod is that they are *mechanically enabled* by this exact dynamic in a way that other forces aren't; the need for land elsewhere is inescapable, the need for electricity from elsewhere is inescapable, the need for information from elsewhere is inescapable, and the impact *on* people elsewhere (totally independent of any direct usage of the technology) is not only tremendous — it's repeated over and over and over again by the most visible Winners of the current 'system'.
Anyway: this was a stellar piece of journalism. Thank you for producing this.
thanks Zac, this is such a kind note! I def agree — it’s kind of remarkable how perfect a symbol data centers are for that chasm — and I still need to read The Unwinding…
"But I'm going to attempt to push past that urge and highlight the most actionable part of all of this, which I think (as you point out) what the entire dynamic reveals about *thymos* — and specifically, Silicon Valley's complete absence of both skill and interest in cultivating it."
This is a really good point to highlight. In some ways they have actively cultivated the exact opposite of the 'thymos' concept in their AI arguments and discussions. I think it's one of the greatest corporate failures of communication in the past 50 years, and there have been a few!
Yes, totally! It presents indeed as a complete failure to internalize such a thing is even necessary - in part IMV because the tech portion of the Bay has become such a transplant city, and in part likewise because dealing with mammoth volumes of money (I mean this non-conspiratorially as myself a participant in this whole world) is just part and parcel of the regular course of business. Both of these I think are exceptionally alienating as compared to the median experience of life:
And I think the 'City Council Meeting' is the one place ordinary people feel the playing field is level enough to push back against this type of massive wealth and resource disparity. Politicians haven known this for eons, Silicon Valley is just figuring it out.
Awesome piece! This is the kind of journalism I come to Substack for.
It's fascinating that we can love talking to ChatGPT but hate the infrastructure that makes that possible. It's almost as if people have decided chatbots are good, AI is bad. It's also really interesting that this protest is uniting people across the political spectrum. Almost as if in the long arc of politics, we're coming back together around a common 'enemy' of sorts.
All politics are local they say, is it possible that datacenters and AI is the spark that takes us back to this after enduring a period where politics have been anything but?
Your point that larger promises can produce deeper suspicion has a revealing counterpart in Ohio. Regulators approved an AEP tariff requiring data centers to pay for at least 85 percent of their contracted capacity, even if they use less, on contracts lasting at least 12 years. After approval, AEP's large-load forecast fell by about half, suggesting that much of the earlier projected demand had not represented firm commitments.
As your reporting highlights, data centers are increasingly discussed as though their only function were to support AI. Often overlooked is their essential role in modern healthcare, including the secure storage, processing, and timely retrieval of electronic health information.
How many centers are ultimately needed or hoped for across the U.S. and on the planet? Is there a map? As long as the groundwater isn't contaminated or used up and the noise pollution is contained, what's the big deal? I stopped twice while reading this to ask Chat what was meant by this or that! Happy Birthday, Jasmine! My oldest childhood friend turns 68 tomorrow.
This reads like a classic illustratation of Sheila Jasanoff's "civic epistemology:" the culturally embedded ways the public decides which knowledge claims, and which claimants, deserve trust. The NDAs, expedited reviews, redacted power estimates all serve to provide a clear answer to "does this actor behave like someone who can be held to account?" which is a more significant question than simply "are these numbers right?." Treating skepticism as ignorance or a problem of misinformation is to fail to recognize this.
I think Sheila's "technologies of humility" are very relevant here, the questions in need of an answer are: who will be hurt if assumptions are wrong? who benefits? and how can we know?
Quite an interesting trip, although I'm sure this is only one path through the US explored. Give me a shout if you're doing something similar in the UK.
Applications are growing like a rash in the UK, often in very ecologically sensitive areas, we need more long term impact data so that we can make good decisions, but are not getting the opportunity
So what I mainly want to do is just post a string of like 52 confetti emojis, because this was definitely the best single piece of reporting on data centers I've seen anywhere. But I'm going to attempt to push past that urge and highlight the most actionable part of all of this, which I think (as you point out) what the entire dynamic reveals about *thymos* — and specifically, Silicon Valley's complete absence of both skill and interest in cultivating it. I don't think that phenomenon is by any means exclusive to SF; in many ways, we're simply entering the third decade of the 'Unwinding' Packer talks about as we've sorted and polarized the macrosocial/macroeconomic forces that shape America out and away from most of America. The Bay is simply one such force.
But what I think causes data centers to be such a irresistible lightning rod is that they are *mechanically enabled* by this exact dynamic in a way that other forces aren't; the need for land elsewhere is inescapable, the need for electricity from elsewhere is inescapable, the need for information from elsewhere is inescapable, and the impact *on* people elsewhere (totally independent of any direct usage of the technology) is not only tremendous — it's repeated over and over and over again by the most visible Winners of the current 'system'.
Anyway: this was a stellar piece of journalism. Thank you for producing this.
thanks Zac, this is such a kind note! I def agree — it’s kind of remarkable how perfect a symbol data centers are for that chasm — and I still need to read The Unwinding…
"But I'm going to attempt to push past that urge and highlight the most actionable part of all of this, which I think (as you point out) what the entire dynamic reveals about *thymos* — and specifically, Silicon Valley's complete absence of both skill and interest in cultivating it."
This is a really good point to highlight. In some ways they have actively cultivated the exact opposite of the 'thymos' concept in their AI arguments and discussions. I think it's one of the greatest corporate failures of communication in the past 50 years, and there have been a few!
Yes, totally! It presents indeed as a complete failure to internalize such a thing is even necessary - in part IMV because the tech portion of the Bay has become such a transplant city, and in part likewise because dealing with mammoth volumes of money (I mean this non-conspiratorially as myself a participant in this whole world) is just part and parcel of the regular course of business. Both of these I think are exceptionally alienating as compared to the median experience of life:
And I think the 'City Council Meeting' is the one place ordinary people feel the playing field is level enough to push back against this type of massive wealth and resource disparity. Politicians haven known this for eons, Silicon Valley is just figuring it out.
Awesome piece! This is the kind of journalism I come to Substack for.
It's fascinating that we can love talking to ChatGPT but hate the infrastructure that makes that possible. It's almost as if people have decided chatbots are good, AI is bad. It's also really interesting that this protest is uniting people across the political spectrum. Almost as if in the long arc of politics, we're coming back together around a common 'enemy' of sorts.
All politics are local they say, is it possible that datacenters and AI is the spark that takes us back to this after enduring a period where politics have been anything but?
Phew! Quite a read but now I know a little bit more . Thank you
Your point that larger promises can produce deeper suspicion has a revealing counterpart in Ohio. Regulators approved an AEP tariff requiring data centers to pay for at least 85 percent of their contracted capacity, even if they use less, on contracts lasting at least 12 years. After approval, AEP's large-load forecast fell by about half, suggesting that much of the earlier projected demand had not represented firm commitments.
Great piece.
As your reporting highlights, data centers are increasingly discussed as though their only function were to support AI. Often overlooked is their essential role in modern healthcare, including the secure storage, processing, and timely retrieval of electronic health information.
67 decibals may be underselling the harm. 67 decibals 24-7 is getting really close to the 70 limit that has been said:
However, 70 decibels also represents the upper limit of the EPA recommended 24-hour average
noise level exposure. Even if it’s not considered a dangerous level, prolonged exposure to noise
levels above 70 dB can still lead to hearing damage or hearing loss.
source: https://mgaleg.maryland.gov/cmte_testimony/2024/ecm/1I7X18LiLthdHKhxoqw3LAFT_0-LwKMXh.pdf
This is an amazing article, Jasmine. Kudos for taking the trip and for the excellent writing.
This is why I am HERE on Substack. We can no longer get this quality of narrative about an ostensibly complex and divisive topic.
Definitely restacking!
How many centers are ultimately needed or hoped for across the U.S. and on the planet? Is there a map? As long as the groundwater isn't contaminated or used up and the noise pollution is contained, what's the big deal? I stopped twice while reading this to ask Chat what was meant by this or that! Happy Birthday, Jasmine! My oldest childhood friend turns 68 tomorrow.
This reads like a classic illustratation of Sheila Jasanoff's "civic epistemology:" the culturally embedded ways the public decides which knowledge claims, and which claimants, deserve trust. The NDAs, expedited reviews, redacted power estimates all serve to provide a clear answer to "does this actor behave like someone who can be held to account?" which is a more significant question than simply "are these numbers right?." Treating skepticism as ignorance or a problem of misinformation is to fail to recognize this.
I think Sheila's "technologies of humility" are very relevant here, the questions in need of an answer are: who will be hurt if assumptions are wrong? who benefits? and how can we know?
Jasmine, I feel like your reporting on data centers is the missing axle for the 2026 midterms. This was a rich and informative read, thank you.
Quite an interesting trip, although I'm sure this is only one path through the US explored. Give me a shout if you're doing something similar in the UK.
Applications are growing like a rash in the UK, often in very ecologically sensitive areas, we need more long term impact data so that we can make good decisions, but are not getting the opportunity
WE KNOW WE NEED TO DROP THE IDEA OF ANY DATA CENTER ..MAYBE THE OCEAN