The AI backlash, like a highly evolved virus, is now endemic. Anti-tech hysteria was perhaps once quarantined to bucolic villages and “5G causes cancer” hippie communes built around NIMBYism. Now, there is a coherent consensus from both left and increasingly the right that communities must do everything they can to halt the progress of data centers and return to some pastoral era before tech’s domination. Where we’re going, no one seems to know.
First it was New York in mid-July, with governor Kathy Hochul bowing to progressive activists to sign a one-year statewide ban on new data center construction while New York assesses the environmental impact of these facilities. A week ago, it was Texas governor Greg Abbott — no progressive — who decreed that data centers in the Lone Star state must be “audited” for their energy consumption before they can proceed (cue the cult audit as depicted in Paul Thomas Anderson’s The Master). The timeline for completing such a procedure has not been disclosed.
The backlash against AI and specifically data centers seems to mostly be raw nerves, inflamed by growing fears of affordability and the concurrent rise in American inequality. Abbott’s pro-growth agenda was felled by consumer anger over high energy prices; Texas has seen prices rise 40% since 2020. That’s the same trigger nationwide, accompanied by a variety of other movements and concerns bolted on for good measure.
Yet, populist fears over power grids and water supplies seem to belie a much more robust hostility to future growth.
A few weeks ago, Jasmine Sun reported a nice travelogue on the AI backlash from Wisconsin (I recommended it last week in Securities). She focused part of the story on Janesville, former Speaker of the House Paul Ryan’s home base that gained fame through an eponymous book that retold the rise and fall of General Motors and its massive facility at the heart of the city. Since GM departed in the aftermath of the Great Financial Crisis, the abandoned facility has remained a polluted wreck for almost two decades.
Sun interviewed activists fighting against a proposed data center that would rebuild that land and turn it into an economically productive base. Admittedly, the project won’t bring back the wealth and jobs of an auto factory, but it’s the best plan the city has seen in decades. As a local booster said to her, “We will have a data center down there, or we will have a giant slab of concrete. The numbers make no sense to anybody other than this industry.”
The activists are nevertheless undaunted and fully opposed. As Sun wrote, “I ask the three what they envision being built here instead, but don’t get much of an answer. I suspect that [the booster] is partly right—maybe this is a choice between a hyperscale data center and a concrete slab. And Alyssa, Cathy, and Kaycee would rather pick the slab.”
That lust for stagnation electrifies the AI backlash. There is no local tolerance for future risk. These activists don’t want their communities to be caught up in a tech bubble that leads to a corrugated metal husk on top of a desolate concrete slab for another few decades. In Janesville in particular, those fears are exacerbated by the searing experience of Foxconn’s failed $10 billion factory in Wisconsin, which I chronicled for TechCrunch back in the day. In some communities, today’s hopes have turned to tomorrow’s disappointments for far too many cycles.
Sitting here ensconced at a venture firm, I try to take a breath and reflect on one of the most impactful books I’ve read in the last few years, The Weaponization of Expertise by Jacob Hale Russell and Dennis Patterson. They argue that the populists are actually sitting quite pretty given the elite stupidity of recent times. “We argue here,” the two write, “that what most unites populism—both left and right leaning—is neither their political agenda nor a set of attitudes such as racism but rather their skeptical perspective toward policy expertise.” They pithily summarize their thesis as “expert overreach actuates populist skepticism.”
Reading their work, one could have sympathy for the Janesville activists and others of their ilk. They look at the jet-setting-into-flyover-state execs from Silicon Valley as aliens from another world, who promise in their pitch decks to completely revolutionize these overlooked communities just as they promise their VC backers exponential returns. The power law of startups that make venture portfolios profitable doesn’t apply to these communities, who each may attract a single company and thus lack the diversification to secure a winner. Thus, members of the public are to be avoided at all costs, NDAs are to be widely distributed and confidentially signed, and zoning permits must be issued without transparency or public comment, all in a bid for technocratic efficiency.
In the duo’s telling, the populists see through this charade; after all, they’ve watched this story before and seen its ending. So they connect with each other to stop a wasteful elite project before it gets underway and deepens the penury of their local community. As they write about how populists view elites, “Technocrats have historically, after all, not been particularly effective in delivering on their promises, and their help may not be necessary.” Leave it to us, say the populists. Also, get the f’ out.
There’s just one problem to this narrative, and it’s the future of that concrete slab.
I could understand the backlash and the fight if the activists had an alternative plan for Janesville’s economic development. They don’t. In fact, no one really does. That’s why there is an enthralling, award-winning book on the riches-to-rags town, about the impossibility of rebuilding the engine after the engine builder leaves. Russell and Patterson are right that elite hubris and its most vile form — condescension — is the ultimate trigger of populism. Yet, infantile rebellion and shouting “no” is hardly any more useful.
The populists are clearly gaining ground, as the statewide moratoriums from left and right attest. I am struck by the inability of leaders to do much of anything besides repeatedly hit the pause button in the face of such widespread anger. Elite inaction amidst populist acrimony is the singular theme of our era against a monstrous and growing inbox of pressing challenges. To stay competitive in a world of change requires ceaseless but balanced risk-taking, building a portfolio of initiatives to be ready for whatever might come. The hyperscalers are investing upwards of a trillion dollars — shouldn’t every community find some way, any way, of getting a slice?
I often get asked about Korea’s industrial policy, and why it has remained so successful for so long. The answer is it just keeps getting on with it. This week, the country announced its “Seven Major SEED” initiative, which includes “commercial deployment of indigenous small modular reactors (SMRs) by 2035, a moon landing by 2030, development of a domestic 100-qubit quantum computer by 2029 and commercialisation of brain-computer interface products by 2035.” Which of these initiatives will pan out? Here’s a hint: the bureaucrats in Seoul have no more idea than any venture capitalist. That’s why you fund them all, experiment, see what happens and double down on what works.
Russell and Patterson’s elite condescension is really just a vituperative framing of an important skill that’s sorely lacking: leadership. Leaders, at least as I define them, need to immerse themselves in the future, assess coming risks and opportunities, and deploy everything in their wherewithal to accomplish whatever needs to happen to maximize the goals and values they and their flock have set forth. Populists don’t trust leaders, and so we get a concrete slab that will soon be eligible to be placed on the National Register of Historic Places.
Populists don’t trust their leaders, and increasingly, leaders respond in kind. I don’t know how to break that cycle of decline. Yet I do know that stagnation is not a business plan, nor does it provide the esprit de corps required for a nation to be competitive this century. When everyone is saying “no,” it’s hard and possibly even impossible to say, “yes.” That’s the virus at work, taking the determined optimism of future generations and infecting it first with doubt, then with anger, and then with abject bitterness. There’s a better approach to the aimless and directionless, and it’s this: how about forward?
A side note on defense
Stagnation is hardly the province of economic development alone. Swing over to the Pentagon, where the Wall Street Journal reported this week that the United States and Ukraine fought a simulated drone war that the Americans predictably and comprehensively lost:
The exercising troops sent into battle heavy armored vehicles that threw up dust and were easy for Ukrainian reconnaissance drones to spot, according to the participant. Those drones were followed by drones dropping explosives and first-person-view drones, which smash into targets in combat, this person said.
[…]
Ukrainian drones were doing this so quickly with U.S. vehicles that they had to be sent back into the exercises as new units, or “re-spawned,” to keep the exercise going, a U.S. official briefed on the drill said.
It is obvious that war has radically changed thanks to autonomy. Old strategies are dead, and need to be burned in a forest fire to lay the seeds for a new generation of ideas by whomever becomes this generation’s Alfred Thayer Mahan. We are making changes to both budget and strategy, just slowly. Too slowly to really soften the blow of the next crisis, but slow enough that everyone can feel comfortable that nothing fundamental really has to change.
Also check out
Why everyone in tech is so sad, and what AI is going to do about it.
A voyage to the edge of the world.
How to predict the future when the past is absurd.
The boomers ruined Europe; Harry Potter almost ruined the grid.
The rise and fall of Carlos the Jackal, the notorious terrorist-for-hire.
Silicon Valley and Philip K Dick, a love story?
There is an economic calculation rationale why data centres are so susceptible to NIMBYing. When I worked with data centre developers, helping them in conversations with the UK government, the overwhelming argument is that the data centres’ economic impact is pretty low. In this context, economic impact = economic multiplier to a pound / dollar / yen of investment.
Economic impact can be calculated using any investment’s capex and opex: in capex, we have the building materials (predominantly non-local in this context) and employment from building a site (huge, but one-off). In opex, we have supply chain spend and employment. The last bit is what ultimately trips developers up, because employing people gives a substantial multiplier — people go on to spend their wages on goods and services, like barbers, and then barbers would go on to spend their wages on groceries, etc etc. This spend tends to trickle through the economy substantially. The materials used to build the centres are also largely imported, contributing negatively to the calculation.
And the above is the back of an envelope calculation. Once the appraisals start (appraisal = is this project worth doing, taking into account opportunity cost), the environmental and social impact will be taken into account, and those things also have an intrinsic £ / $ / ¥ value (you might be surprised that someone’s life also has an economic value). Worryingly, I am not aware of any appraisals being done, so it might be that the back-of-the-envelope calculation leaves it dead-on-arrival.
Now the benefit of a data centre transcends borders. So if I’m the UK government, I’d be asking myself why do we need to build in the UK, while a data centre infrastructure in France will use cleaner and cheaper energy (ie nuclear) and we still get compute at home. There’s of course more nuance here, but you can start to see how the NIMBY argument is, economically, hard to overcome.
I imagine the risk of a political backlash, perhaps coupled by a lack of international cooperation, is tilting the decision one way and one way only, even that means letting concrete slabs sit doing nothing with a slim chance an IKEA could be built there in the distant future.