Members of MADSA have been outraged over the environmental, social, labor, and economic harms wrought by data centers, and have been encouraged by the widespread responses against them. We’ve taken some initial steps to spin up a MADSA anti-data center effort that can help further struggles across the state of Wisconsin. This article recounts some of our progress and activity thus far, and outlines some problems and potential strategies to go forward.
MADSA’s Data Center Fightback group began meeting at the end of August, with around two dozen attendees at our initial meeting. Several projects were discussed and begun at those initial meetings, with members seeking to fight against environmental harms, corporate corruption in government, and to take advantage of this ideologically diverse struggle. Some members were motivated by Democratic gubernatorial candidate David Crowley’s lukewarm stance on data centers — he refuses to support statewide actions like moratoriums, saying local communities must decide for themselves, and has argued that data centers have positive economic impacts. His stance on the issue was arguably the greatest divide between him and Francesca Hong, whom he narrowly defeated in the Wisconsin Democratic primary a few weeks before our group began meeting. Republican gubernatorial candidate Tom Tiffany has taken advantage of Crowley’s stance on data centers to proclaim that data centers were the “only difference between Hong and Crowley.” Tiffany has opportunistically taken up the populist issue himself, appropriating and twisting the slogan “America’s Dairyland, not Dataland” coined by the Party for Socialism and Liberation. In other words, Crowley’s rejection of concerns over this issue has aided Tiffany in his efforts to turn this issue from a liberatory socialist one (as our comrades in PSL have led the fight on) into a reactionary fascist one. MADSA members sought to intervene in this through establishing our own anti-data center effort.
At those initial meetings, we also planned a teach-in, which we hosted on September 19. We sought to ground our initiative in more rigorous evidence and a stronger political analysis. We also saw the teach-in as a way to bring together newcomers who may not self-identify as organizers from the outset, but who may be emboldened to act through discussion and education with others upset about data centers. To that end, we text-banked MADSA members and created short-form video content to advertise the event.
At the teach-in, we reviewed and discussed environmental harms, sound pollution, strains on the grid, and more (see our slides here). Many questions were raised and clarified, particularly around the mechanics of some of the environmental factors. Why does closed-loop cooling still use water? Closed-loop systems still generally use evaporative cooling which consumes water, they need to periodically flush water to avoid mineral buildup, and they consume water indirectly through even higher electricity usage. Why is fresh water needed for data center cooling? Fresh water prevents scaling and corrosion, as well as bacteria buildup in warm water. Finally, we discussed several ongoing projects around Wisconsin, tax exemptions and tax incentive districts, and the community struggles that have been successful in blocking data centers thus far.
Image by Nathan Kim, Madison Area DSA. CC BY-NC-SA 4.0.
This empirical overview was necessary for a basic understanding of the fight we were up against, but the event organizers also felt that something more sophisticated was necessary. Environmental harms of data centers are constantly shifting, and more important than specific statistics is to understand the current contradictions within capitalism that give rise to these varied harms. We therefore also presented our own analyses of the political economy of AI and its recent history. We argued that the rise of the cloud economy, a persistent crisis of underinvestment and overaccumulation over the past several decades, and extreme concentration of capital within a few firms have promoted a “throw more data and compute to win first” strategy at AI development over slower forms of scientific research. We also discussed the general mechanisms by which capitalism enforced technological development through competition between capitalists, locating labor discipline at the heart of capitalist adoption of technology. That constituted a self-undermining strategy given the tendency of the rate of profit to fall with greater fixed capital investments.
We also laid out some principles for organizing around data centers from our positions as socialists. Event organizers felt that although we have every right to participate in struggles purely to defend our communities, socialists also had to constantly push to establish more capacious horizons that data center struggles could be part of. We believed that only through the elaboration of that more expansive politics, an articulation that can only come through practice and continual reflection, could we avoid the capture of issues like these from reactionary opportunists like Tiffany and instead make the data center fight part of a broader socialist movement. Of course, it is worth saying that the specific ways that struggles could be expanded could never be predetermined and should never be imposed on by socialists (only reached through collective analysis). But we felt that some baseline principles would nonetheless be useful for us: we should choose existing struggles and attempt to further politicize existing concerns, we should make connections between different movements and issues, and we should share what a real alternative to current conditions would look like. We also discussed the need for developing long-term working-class formations, through a conscious development of institutional infrastructures and merging so-called “self-selecting” and “class-based” issues.
The teach-in was arguably most useful to surface disagreements among participants and for us to understand the challenges ahead of us. One lively disagreement revolved around the role of organized labor within this fight: though AI as a whole is an obvious source of labor displacement, it is also true that data centers in the here and now have provided a massive construction boom due to the weak economy and lack of investment outside of data centers, and are therefore highly prized by trade unions like the IBEW. Some members insisted that labor must avoid such short-sighted approaches and look at the effects of data centers on labor displacement after they are constructed, while others argued that it was a faux pas to argue as much to unions directly, given that construction workers’ livelihoods are so dependent on data center development.
Other disagreements included ideological disunity within participants on AI safety and existential risk; although some are truly concerned about the harms of AI to humanity, many believe it is a mistake to use the language of existential risk and that the topic in general is dangerous to rely on because it can mystify AI. Others raised concerns over fighting data centers in Madison specifically, given that Madison is only home to smaller centers at present like the Dane County Data Exchanges 1 and 2. Still, a potential data center by Cloverleaf on Madison’s East Side and the formation of new zoning frameworks around data centers have raised concern and have motivated additional investigation and planning.
As a result, there was not an immediately obvious way to set in motion the capacious vision we agreed that we needed, but such an approach would only be possible anyhow with greater knowledge on existing fights and how we could bring them together. To that end, we shared contacts and made plans to set up meetings with several organizations and community efforts brought up during the teach-in.
If you’d like to join us in this fight, watch for upcoming events on the MADSA website, or please join the #data-center-fightback channel on the MADSA Slack.
Leave a Reply