Grandma Can't Outbid
the Cloud

Some communities are organized around throughput. Some communities are organized around presence. The communities organized around presence — where elders can stay, where proximity between generations is protected, where the slow infrastructure of relationship is treated as load bearing rather than as friction — are the communities that make care possible at all.

This brief is about what happens to those communities when compute infrastructure arrives. It is not an argument against data centers. It is an argument for a kind of place that compute economics, in its current form, is actively dismantling without ever being asked to account for the dismantling.

The argument extends a position developed in the Joy Repair working paper Economy Runs on Love, which holds that love is not outside the economy. Love is the precondition that allows economies to survive their own violence. If relational infrastructure is the substrate that keeps economies survivable, then any development pattern that systematically degrades it is producing a structural cost the conventional accounting of growth does not capture. The data center boom is producing exactly that cost, at exactly that scale, in exactly the places least equipped to name it.

The cloud is not in the sky. It is sited. Every data center lands in somebody's county, draws on somebody's grid, draws down somebody's water, and reprices somebody's land. The decision about where it lands is also a decision about whose care infrastructure will be extracted to power somebody else's compute.

That decision is currently being made in hundreds of counties across the United States, in rooms most residents will never enter, on timelines too compressed for public deliberation, and in language too technical for the people whose neighborhoods will be reorganized by the outcome. The deals are usually announced after they are functionally complete. The tax abatements are usually granted before the public sees the line items. The rate increases arrive on bills sometime later, after the megawatts have already been promised.

fig. A — The Grandmother Walk Map · Before and After

Grandma's walking world (green) vs. the data center's impact radius (red). Watch what happens when compute arrives.

The structural account of why these rooms systematically exclude the people most affected is developed in the Joy Repair working paper on the Shame Buffer. Consultation increasingly operates as a civic anesthetic. Residents are invited to speak precisely where they cannot meaningfully intervene. The data center approval process is a textbook case. The vocabulary, the timelines, the meeting schedules, the technical thresholds for participation, all of it is designed to absorb dissent without redistributing decision authority. Comment is not the same as deliberation. Public hearing is not the same as public authorship.

Data centers need powered land, water rights, utility upgrades, transmission capacity, tax incentives, and fast approvals. Compute based development optimizes for throughput, scale, energy concentration, and capital returns.

Care based development needs stable housing, intergenerational proximity, public space, child care that costs less than rent, elder services within walking distance, mutual aid networks that took decades to form, and slow trust. Care based development optimizes for belonging, memory, presence, and relational density.

The market can price megawatts. It cannot price a grandmother walking a child home from school.

The non-parallelism of these two development logics is a bifurcation problem, not a tradeoff problem. The Joy Repair 6/7 Threshold Model applies bifurcation theory to social transformation, holding that threshold events are often misrecognized because the transition feels administrative while the consequences become civilizational. Counties do not gradually become a slightly thinner version of themselves under compute pressure. They cross a line, after which relational density collapses faster than institutional repair can respond.

The dominant frame for what happens to communities under this pressure is the language of deserts. Childcare deserts. Pharmacy deserts. Food deserts. Senior service deserts. The desert metaphor describes absence as if it were geography. As if the absence of childcare were like the absence of rain. As if no one were responsible. As if the desert had simply formed.

Geography does not move. Capital does. When a region's land, water, megawatts, zoning attention, and tax abatements are systematically directed toward compute infrastructure — and when the labor of mothers is simultaneously extracted to keep households solvent under rising utility costs and rents — desert is the wrong noun. The accurate noun is extraction center.

The decision to rename is not rhetorical. The Joy Repair propagation work argues directly that accurate naming is not rhetoric. It is the first condition of collective self defense. Words like desert do not just describe inaccurately. They obscure agency, distribute blame onto absent actors or onto the affected populations themselves, and make accurate response politically impossible. The terms introduced here are accurate. The discomfort they produce is the discomfort of seeing the mechanism plainly for the first time.

The third tier is the one that gives the second tier teeth. Care based development sounds like a slogan. Elder centered place based stability sounds like something a county council could plan for, fund, zone for, and measure. It is also the tier the compute based model most directly degrades.

The argument here is not sentimental. It is structural.

Elders are the population for whom proximity, memory, and slow trust are the substrate of daily functioning. They cannot easily relocate. Many of them are on fixed incomes that cannot absorb the residential rate increases compute infrastructure produces. They depend on networks of relationship that took decades to build. They hold the institutional memory of a place. Who is related to whom, which corner used to be what, which family carried which grief, which neighbor will check on whom during a heat wave.

When a community pushes elders out — through rate increases, housing displacement, service desertification, mobile home park sell-offs, the relocation of senior services to other counties — it does not just lose its elders. It loses its capacity to be a place. The compute based development model treats this as friction. Care based development recognizes it as the foundation.

The seven generation framing that grounds this argument is developed across the Joy Repair body of work, most directly in Punk Rock Is Love. A community that loses its elders to megawatt economics has lost the population that makes seven generation thinking even possible to attempt.

This is an exploratory pass. The goal is pattern detection, not yet causal proof. The question we are asking each community is the same: What was the place organized around, and what came under pressure when compute infrastructure was proposed or built?

fig. C — Five Communities · One Pattern

Five real places. Five different stages of the same extraction. Click any card to read the story.

Pattern detection, not causal proof. Five geographies, one mechanism.

The mechanism that makes these approvals possible is documented in the Joy Repair working paper Civic Apparatus, which argues that modern civic systems often mistake stakeholder circulation for democratic control. Each of the five communities profiled here has been processed through some version of that confusion. This is not a series of unrelated local failures. It is a single governance pattern operating at national scale.

The most common defense of compute based development cites state level averages. State level studies find no statistically significant relationship between the number of data centers in a state and that state's electricity prices. State level studies are correct, on their own terms.

Care does not happen at the state level. Care happens at the block level. Care happens at the kitchen table. Care happens on the walk between the school and the house where the grandmother lives.

fig. B — Megawatt Load vs. Care Fragility · Northern Virginia · 2018–2025

As data center megawatt load rises, residential utility rates and elder disconnection rates rise with it. The lines cross. That is the threshold.

The Joy Repair Conditions of Seeing framework names the problem directly: care does not occur at the scale most institutions measure. It occurs at the scale of blocks, kitchens, rides home, and repeated human contact. The same dataset that produces no state level signal produces a 267% wholesale price increase at specific grid nodes near data center clusters. State averages mask local effects. Northern Virginia's pressure gets diluted when blended with rural Appalachia.

Before approving any compute infrastructure project, every community should be required to publish answers to the following questions. The audit is organized by life cycle care infrastructure, with elders and place based stability foregrounded.

The proposal here is not radical. Environmental Impact Statements have been required for federal actions since 1970. Care infrastructure deserves at minimum the accounting that environmental infrastructure already gets. The fact that it currently does not is the entire problem this brief is naming.

Not proof of causation, yet. Just pattern detection. We are asking whether communities are trading care capacity for compute capacity without naming the trade. The five community survey above suggests they are. The pattern is consistent enough across very different geographies — coastal Maine, west Texas, northeastern Pennsylvania, Northern Virginia, rural Louisiana — to justify the full investigation.

The full investigation would include geographic information systems analysis of land values, zoning decisions, tax abatements, utility rate cases, energy demand projections, child care availability data, senior service location data, multigenerational household data, and housing displacement metrics. All measured at the block group or census tract level rather than the state level. It would produce a Place Based Stability Index that local governments could use as a planning tool.

Before approving AI infrastructure, every community should be required to ask three questions.

A community that cannot name what it is organized around cannot make an informed decision about what to consume in pursuit of growth. It can only consent to extraction it does not yet have language for.

The affirmative vision is not theoretical. It has been built and held, in specific places, by specific people, within living memory. The Joy Repair documentation of the Indianapolis Deaf Club, of Dance Is Music for the Deaf, of the Zawadi Exchange in Mapleton-Fall Creek, all of it documents what elder centered place based stability looks like when it works. The communities most capable of surviving technological acceleration will not necessarily be the most computationally advanced. They will be the ones that preserved the ability to belong to one another.

A full Joy Repair Special Report on Compute Based Development, Care Based Development, and Elder Centered Place Based Stability is warranted. The exploratory pass is sufficient to establish the pattern. The full investigation would include the following.

The brief and the report together would constitute the first sustained argument that elder centered place based stability is the affirmative ground on which the data center conversation should be conducted. Not as opposition to compute infrastructure, but as a precondition for any community making an informed decision about it.