Too Much Computer: Ingrid Burrington discusses Hyperscale with Paris Marx
Oct 22, 2026 · 1h 0m
More Technology podcasts
Summary
Ingrid Puig interviews Harris about his new book, *Hyperscale*, which examines the unsustainable environmental and social costs of the massive data center build-out driven by AI. The conversation explores the financialization of cloud computing, the role of venture capital in enabling this infrastructure, and the disconnect between tech narratives and local community impacts. They also discuss the history of personal computing, the rise of AWS, and the use of greenwashing tactics like water replenishment markets to obscure the true resource intensity of these facilities.
Topics discussed
AI investor profits and upcoming IPOs
Announcement of the book 'Hyperscale' release
Public backlash against data center construction
Upcoming book tour dates and events
Introduction of guest Ingrid Burrington
Guest background and previous works
Why the podcast content became a book
Choosing data centers over Elon Musk as a topic
The temporal advantage of books over podcasts
Challenges of writing a long-form book
Organizing the book's structure and history
Explaining data center basics and issues
Addressing common bad-faith arguments
Debunking water usage controversies
Validating water usage concerns in data centers
Global precedents for local data center opposition
Scale issues and exceptional cases in sustainability
Comparing data centers to automobile dependency
The push for computationally intensive AI tools
Community-level impacts and solidarity movements
Grid, water, and neighborhood strain from data centers
Debating the need for sustainable computation
Early personal computing and accidental openness
Liberatory seeds in early tech history
The financial project behind cheap computation
Workplace automation and white-collar computing
Less commercialization in the early web era
Hidden hardware production behind tech products
Omission of production in libertarian tech narratives
Nostalgia and the incomplete picture of the 90s
Labor exploitation in early Apple manufacturing
Silicon Valley as a specific geography
ZERP and the rise of cloud computing
Venture capital glut enabling AWS adoption
Risk investment and market cornering strategies
Recouping costs through acquisitions and exits
Mergers in online payments (Confinity and X)
Amazon's strategy of domination before profit
Outsourcing server management to the cloud
Security benefits and risk offloading in cloud
The elusive promise of cloud cost savings
Concentration of compute for AI model training
Palantir and the role of centralized data
Marketizing and financializing server assets
Water replenishment markets and offsets
Critique of market-based climate solutions
Fossil fuel use and corporate image management
Using credits to negate actual resource usage
Effectiveness and longevity of carbon offsets
Risks to offset projects (wildfires, logging)
Local water needs vs. long-termist thinking
Accountability for emissions and climate targets
Circular financing and government incentives
Sovereign AI and state-led data center builds
Replicating US tech models for national security
Long-term investment vs. immediate returns
Constraint and the myth of unlimited compute
Moving beyond shaming for tech change
Convenience and the difficulty of leaving platforms
Addictive design and social network lock-in
User interface design and behavioral change (taxis)
Expectations for user effort in tech adoption
Hope in current public anger at tech
Redirecting anger toward a social tech vision
Democratic co-creation of technology
Humorous take on data center awareness
Reflection on a decade of explaining data centers
Final thoughts on the book's framing
Recommendation of 'Networks of New York'
Show credits and support information
More from Tech Won't Save Us
Listen ad-free on Castria