Should we let AI infrastructure be built at 'rocket speed' before the grid is ready?
SpaceX brought the largest AI cluster in the country online in months on rented turbines and portable chillers, then spent the summer with outages, an executive exodus, and a lawsuit from neighbors. Is building compute faster than the surrounding energy and cooling systems can support a necessary price of progress, or a shortcut that shifts the cost onto tenants, communities, and the grid?
Commentaires (1)
In today’s episode of Minds, Bodies, and Terawatts, dated September 3, 2026, we dug into The Information’s report that SpaceX handed its Memphis-area AI data centers to Starlink’s engineering chief after months without backup power or cooling, missed uptime targets, and interrupted model training for paying tenants. The episode argues the discount SpaceX bragged about was really the missing safety net, and that Musk’s decision to cast his own gas turbine blades in Texas is the moment a compute landlord discovers the grid, not the chip, is the constraint. With Google’s $920 million-a-month contract starting at full rate in October, the bill for speed is arriving all at once. Listen to the full episode and tell us where you land: is this the messy but inevitable way infrastructure gets built, or a warning about what happens when AI outgrows the human systems around it?
Related reading on unscarcity.ai:
Envie d'aller plus loin ?
Obtenez le plan complet dans <em>L'ère de la post-pénurie : Repenser la société à l'ère des machines</em>
