AI on the Ballot (Stratechery Article 11-4-2026)
1. Key Themes
Theme: Political opposition has become the binding constraint on AI buildout
The constraint on AI has migrated from compute to memory to power to politics
Thompson argues the bottleneck keeps shifting, and the final one is the hardest to solve because it interferes with fixing the others.
"First we were short on compute, but that ramped; now we are short on memory, but that will ramp; we were coming close to being short on power, but that will be rectified. The final constraint is political, and that may be the most imposing constraint of all. Not only is it harder to solve, but it actually interferes with solving the other constraints."
Data center opposition is bipartisan and electorally potent
This is not a partisan pendulum that swings back in two years. Both parties treat opposition as a winner, even in the state that benefits most from the buildout.
"Both parties see opposition to data centers as a winning cause, and not just nationally: that thread has multiple examples of local officials straight up recalled for approving data centers in their localities."
"Republican incumbent Greg Abbott, I would note, already declared a moratorium on new data-center permits in September; that's how much of a hot-button issue this is, including in the state that is the biggest opportunity and beneficiary of the data center buildout."
Local ballot measures are going one way
"...nearly every relevant local ballot measure opposing data centers won or is winning (the only exception so far was in North Dakota)..."
Theme: Scarcity may last far longer than commodity-AI theories assume
AI may not become a commodity
"Many of the theories about AI assume that it becomes a commodity; I think it's just as likely we are looking at scarcity for much longer than we expected, which will have ripple effects on everything."
Constrained compute shifts value to efficiency, favoring Nvidia's tokens-per-megawatt pitch
"A world without power or data center constraints is one where it is, in the very long run, very difficult to maintain Nvidia's margins; once those constraints start to bite, however, an advantage in tokens/megawatt is worth a lot of money."
Free-compute consumer services may not repeat in AI
"...what they will miss out on is the equivalent of the Web 2.0 wave of free Internet services that were predicated on compute being free. Now there's a good chance that AI is going to have meaningful marginal costs for a long time to come."
Theme: A constrained world reshuffles winners and losers across the AI stack
Hyperscalers gain share and cover to moderate capex
"The downside for the neoclouds is upside for the hyperscalers, who are more likely to grow their share of AI compute than they were before. Moreover, investors may be pleased if the hyperscalers' ambitions are somewhat curtailed; political opposition is an excuse to moderate spending, knowing that everyone else is in the same boat."
Neoclouds face a double squeeze of capital and permission
"It was already getting harder to get capital, relative to the hyperscalers; now you have to layer on that it's going to be harder to get permission, too."
Frontier labs get a pacing mechanism
"Now, however, it is the public that is effectively demanding that the frontier be paced, which is great for the frontier labs' financial outlook."
Startups lose cheap compute and runway
"In a world with fewer data centers, startups will have less access to cheap compute, even as a slowdown in capabilities both means it takes longer to build compelling products and also gives the frontier labs time to actually build products that secure their ownership of the end user."
Theme: Alternatives to terrestrial buildout gain strategic value (China, space, edge)
China's power advantage may offset its chip disadvantage
"China — which actually builds power generation — is not iso power; they are iso compute: Huawei's goal with the CloudMatrix 384 is to get as much compute as possible, power efficiency be damned."
Space-based data centers become more plausible if earthbound opposition persists
"What seems very plausible in the long run is that the demand for compute ends up being so large that there eventually is nowhere left to build, making the vast expanses of space not just an alternative but in fact the only choice."
Edge inference and memory relief favor Apple
"...no company would be a bigger beneficiary of the world running out of places to put AI chips than Apple."
2. Contrarian Perspectives
The tech industry's biggest culprits may be the biggest beneficiaries of the anti-data-center wave
The obvious read is that frontier labs brought this on themselves and will suffer. Thompson argues the opposite: slower buildout relieves the competition-driven R&D spend that threatened to outrun revenue.
"...the biggest revelation from Anthropic's S-1 was that their inference margins were very high (good!), which meant that their massive compute costs were actually in R&D; this is bad because that R&D spend is governed by competition with the other frontier labs, and absent constraints, that spend looked set to surpass revenue sooner rather than later."
Blocked data centers are partly a silver lining for credit-starved neoclouds, and may de-risk space
Being unable to build is less harmful when financing is unavailable anyway, and permitted sites become scarce assets. Permanent opposition also supports demand durability for space-based compute.
"...being unable to build data centers has a silver lining in that they couldn't get the money to pay for them anyways. Moreover, Neoclouds with facilities that are already permitted now have an asset that is going to be scarcer than expected."
"Moreover, to the extent opposition is permanent, the more this actually reduces the risk of data centers in space, because investors will have more confidence in there being enduring demand."
The anti-AI backlash is rooted in a newly available veto, unlike globalization
Opposition isn't merely ignorance; it exists because people finally have leverage they lacked during globalization.
"...building data centers requires permission, which is to say that people actually have a say...people who didn't have a say in globalization are suddenly finding they do have a say about AI, and it's not a surprise they are expressing their disapproval by blocking data centers."
Slower AI dispersion might be a net good for humanity
"Maybe this is a break we needed; what is increasingly clear is it's the break Americans wanted."
3. Companies Identified
- Description: Dominant AI chip maker
- Why mentioned: Short-term loser from fewer data centers and weakened neoclouds, but constraints make its efficiency advantage more valuable
- Quotes: "Nvidia will almost certainly be a short-term loser. Fewer data centers mean fewer Nvidia chips generally, and a weakened Neocloud set is particularly harmful not only because Neoclouds almost exclusively use Nvidia full-rack systems, but also because Nvidia is back-stopping a huge number of them."
- Description: Frontier AI lab that has filed an S-1
- Why mentioned: Its S-1 revealed high inference margins, with compute costs concentrated in R&D
- Quotes: "...Anthropic's updated S-1, but it strongly affirms my initial analysis" and "the biggest revelation from Anthropic's S-1 was that their inference margins were very high (good!)..."
- Description: Launch company with an S-1 and a data-centers-in-space angle
- Why mentioned: Political opposition to terrestrial data centers is an explicit upside scenario
- Quotes: "When SpaceX released their S-1, I explicitly pointed to political opposition to data centers as one of the company's most significant upside scenarios"
- Description: Consumer hardware and edge-compute leader
- Why mentioned: Edge inference gains importance, AI slowdown helps the previous-paradigm winner, and memory-crunch relief benefits it most
- Quotes: "The most obvious advantage for Apple is that in a world of constrained compute, inference on the edge will be that much more important."
Hyperscalers (Amazon and others)
- Description: Large cloud providers
- Why mentioned: Likely winners, with existing sites to repurpose and more political sway; Amazon's Trainium ASIC discounting is cited
- Quotes: "...Neoclouds...are likely to have much less political sway than the hyperscalers — who already have more sites they can repurpose or rebuild."
Neoclouds
- Description: Specialized GPU cloud providers
- Why mentioned: Face credit tightening and permitting difficulty
- Quotes: "The Neoclouds have been at the forefront of credit concerns..."
- Description: Chinese tech company
- Why mentioned: CloudMatrix 384 shows China's strategy of brute-force compute with abundant power
- Quotes: "The CloudMatrix 384...takes up dramatically more space, uses dramatically more power, is almost certainly dramatically more expensive to make, and is dramatically more brittle than Nvidia's NVL72..."
- Description: Payments company and the issue's sponsor
- Why mentioned: Its SaaS Index shows revenue growth accelerating despite the "SaaSpocalypse" sell-off
- Quotes: "Year-over-year growth is now over 30%, with some of the fastest growth coming from the youngest SaaS companies."
- Description: MLB team
- Why mentioned: Thompson's viewing of Brewers games is how he noticed the volume of data center campaign ads
- Quotes: "my World Champion Milwaukee Brewers — demanded my full attention, which also meant I saw a lot of campaign ads, and by far the most common topic...was data centers."
4. People Identified
- Description: Nvidia CEO
- Why mentioned: Argued at a 2025 keynote that in power-constrained ("iso power") worlds, tokens per megawatt is the metric that matters
- Quotes: "Huang is issuing a challenge to every cloud provider tempted to make an ASIC: do you really think that you can keep up with this, and risk running out of power capacity because your chips aren't efficient enough? Don't even try."
- Description: Newly elected Democratic governor of Texas
- Why mentioned: Pledged a data center construction moratorium in the state with over 25% of planned US capacity
- Quotes: "Hinojosa, Texas' first Latina governor, has pledged to impose a moratorium on data center construction in Texas, which accounts for more than 25% of planned US capacity by 2030."
Greg Abbott
- Description: Republican Texas governor (incumbent, lost)
- Why mentioned: Declared a data-center permit moratorium in September, showing bipartisan opposition
- Quotes: "Republican incumbent Greg Abbott, I would note, already declared a moratorium on new data-center permits in September..."
Donald Trump
- Description: US President
- Why mentioned: The midterm results were described as a rejection of him over the economy and the Iran war
- Quotes: "...a midterm election landslide that amounted to a rejection of President Donald Trump following widespread voter discontent with the economy and the war in Iran."
5. Operating Insights
Pay communities for permission to build
Thompson earlier argued builders and their customers should compensate communities, which at minimum raises terrestrial costs and builds goodwill.
"In that Update I made the case that data center builders — and by extension the companies that use them — should straight up pay people for permission to build data centers in their communities."
Stop assuming the benefit of the doubt; the industry must persuade
"The big difference today is that, unlike the first few waves of tech, the industry is going to need to do some convincing, and break the habit of assuming everything it does will get the benefit of the doubt."
Prioritize efficiency (tokens per watt, right-sized models) and rebuild existing sites
When permits are scarce, extract more output from existing sites and design for efficiency.
"...one way to generate more tokens is to build a new data center; another way is to rebuild a data center with new Nvidia chips."
"...if compute is constrained, and thus more expensive, then that will increase the drive to right-size models for jobs, which is a big argument in favor of open models..."
6. Overlooked Insights
Stripe data suggests the "SaaSpocalypse" was a sentiment event, not a fundamentals event
Nearly a trillion dollars of software value vanished, yet payments data from 72,000 non-AI SaaS businesses shows accelerating revenue.
"Stock prices only tell us what investors think will happen, and not necessarily what is happening...while SaaS stocks were selling off, revenue growth was actually accelerating."
Open models face greater regulatory risk under a Democratic Congress
"Moreover, open models may be at greater regulatory risk, as Democrats seem much more likely to pass laws that ultimately favor the frontier labs."