By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
On with Kara SwisherEverything you’ve been wondering about data centers, explained
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
CHAPTERS
Jump to publisher timestampsChapter summaries
Why data centers suddenly became a political flashpoint
Kara Swisher opens with why data centers—once boring infrastructure—have become a hot-button issue as communities push back on AI-related buildouts. Each panelist describes the personal and political moment they realized the backlash had gone national.
- •Kara frames data centers as a rare bipartisan local issue with growing national consequences
- •A panelist describes being flooded with emails from one Pennsylvania community, prompting the creation of a crowdsourced national map
- •A panelist recounts Utah’s Stratos proposal igniting “torches and pitchforks” political energy
- •Elena Schlossberg explains how NDAs and a secretive Amazon-linked campus catalyzed organizing in Virginia
- •Mayor Andy Schor describes how opposition to a nearby hyperscale proposal reverberated into broader local resistance
How AI changes the scale: training vs. inference and soaring power demand
The conversation clarifies what hyperscale data centers are and why AI changes their footprint. A speaker distinguishes training facilities from inference facilities and explains why the new wave is orders of magnitude larger and more energy-intensive.
- •AI model training demands far more compute than prior eras of cloud data centers
- •Two categories are described: training centers (massive loads) and inference centers (latency-sensitive, closer to population)
- •Typical sizes have shifted from tens of megawatts to hundreds of megawatts and even gigawatt-scale proposals
- •Bigger facilities mean bigger energy consumption and broader siting impacts
Secrecy, shell companies, and NDAs: how projects stay hidden until it’s too late
Kara asks why so many data center proposals arrive with little public notice. A panelist describes common patterns seen across thousands of community submissions: shell entities, NDAs, and residents feeling blindsided—often to avoid early opposition.
- •A speaker reports thousands of community submissions alleging secrecy and shell corporations
- •NDAs are described as widespread and confusing to the public, especially when officials sign them
- •The secrecy is attributed to avoiding early public opposition and speeding approvals
- •Kara raises whether security or competition is the justification; a speaker emphasizes opposition avoidance
- •Virginia examples include officials reportedly unable to say “Amazon” due to NDA terms
Virginia’s Digital Gateway fight: disclosure gaps, water risk, and grid instability
Elena Schlossberg explains why Virginia has become a cautionary tale, focusing on the Digital Gateway controversy and what she sees as systemic non-disclosure. She emphasizes not only water and land-use impacts, but grid reliability risks that extend beyond any single neighborhood.
- •Elena argues state tax incentives shift costs to residents and encourage overbuilding
- •She says projects often lack required disclosure on water quantity/quality and power needs
- •She links impervious surface buildout to reduced groundwater recharge and polluted runoff in watersheds
- •She highlights grid risk, describing a major power event tied to data centers disconnecting from the grid
- •She frames the issue as national: reliability, affordability, and oversight affecting all ratepayers
A “small” data center done differently—and why it still got rejected in Lansing
Mayor Andy Schor details Lansing’s attempt to support a downtown-edge data center with strict conditions and no tax incentives. Despite requirements on water use, location, and community benefits, the project faced opposition and ultimately did not move forward.
- •The Lansing proposal was described as cloud storage rather than AI-related
- •Schor says the city required closed-loop cooling and avoided drawing from local rivers/aquifers
- •He describes siting as urban infill on empty lots near infrastructure, with noise ordinance compliance
- •The city rejected tax incentives and required payments toward community energy assistance (“Pennies for Power”)
- •Schor worries rejection pushes projects to places with weaker requirements, worsening environmental outcomes
Cooling and water: why “closed loop” is disputed and why off-site water matters too
The panel debates whether closed-loop cooling solves water problems. Elena challenges the idea that closed loop is a panacea, while another panelist and Logan emphasize that water impacts also come from the power plants supplying electricity, not just the data center itself.
- •Elena argues closed-loop systems still require flushing and can involve chemicals that end up as pollution
- •She questions backup generation practices and raises concerns about diesel generators more broadly
- •A panelist calls closed-loop an “illusion,” arguing it can increase off-site water use via added electricity demand
- •Logan explains there are two water stories: on-site data center use and water used for energy generation
- •The discussion stresses that developer messaging often focuses only on on-site cooling water
Stratos in Utah: natural gas buildout, massive water use, and pollution trade-offs
Logan walks through Utah’s proposed Stratos project and explains how missing details forced analysts to infer likely technologies. He compares combined-cycle turbines versus reciprocating engines, highlighting stark trade-offs between water consumption and air pollution.
- •Logan describes the proposal as a massive natural-gas-powered data center plan with limited public detail
- •He estimates combined-cycle turbines could drive extremely large water consumption due to evaporative cooling
- •He says turbine backlogs push projects toward reciprocating engines, which use less water but emit more NOx pollution
- •He links the Utah siting near the Great Salt Lake to heightened sensitivity during drought and conservation efforts
- •The “free pass” perception—public conservation demands vs. project scale—helped fuel backlash
Construction-phase impacts: turbidity, low pressure, outages, noise, and habitat loss
Kara asks about reports of water quality and pressure problems near build sites. A panelist describes an “arc” of impacts from proposal secrecy to construction disruption and then to operational complaints like constant noise.
- •Reports include increased turbidity/cloudiness and loss of water pressure near construction
- •Some residents reportedly need to re-drill wells due to lowered water levels
- •Electrical issues are described, including flickering and outages during buildout
- •Construction is portrayed as highly disruptive: tree removal and habitat destruction
- •Once operational, complaints shift heavily to persistent 24/7 noise
Who should pay for AI infrastructure? Taxes, incentives, and cost allocation battles
The panel debates whether energy-consumption taxes and incentives meaningfully address public anger. Elena argues removing subsidies and making companies pay their own costs would drive innovation, while Schor supports mechanisms that prevent residents from subsidizing profitable firms.
- •Kara cites Virginia’s energy consumption tax and asks whether it changes outcomes
- •Elena says the tax won’t matter much compared to removing incentives and fixing cost allocation
- •Elena uses her CarMax analogy: if companies pay, they’ll choose cheaper, more efficient options
- •Schor says residents fear rate hikes for profitable businesses and supports policies to avoid cost-shifting
- •He describes Lansing’s refusal to offer incentives and requirement of community-benefit contributions
Grid stress, emissions, and the risk of building for a bubble
Kara turns to power-demand projections and asks Logan to contextualize grid strain and emissions. Logan raises the possibility of an AI hype cycle, but says if the load materializes it would require massive generation and transmission investment—while a Utah facility could dramatically raise state emissions.
- •Logan says the scale of projected load growth would drive a huge buildout cycle for the energy system
- •He suggests efficiency gains could reduce growth, citing past periods where efficiency curbed demand
- •He argues data centers could fund grid upgrades if done transparently with proper tariffs
- •He estimates a large gas-powered facility could significantly increase Utah’s total CO2 emissions
- •He flags upstream methane leakage as an additional climate concern beyond combustion emissions
Guardrails and legislation: demand flexibility, clean power, and public accountability
A listener question prompts the panel to propose guardrails at local, state, and federal levels. The answers focus on demand flexibility, paying for infrastructure, clean energy pairing, transparency, and the political consequences when voters feel ignored.
- •A panelist calls the buildout a “runaway train” and emphasizes zoning/permitting oversight and transparency
- •Logan proposes demand flexibility so data centers reduce load during peak grid stress periods
- •He argues data centers should bring clean energy and storage and pay for broader grid upgrades
- •Schor lists guardrails: water, electricity, maintenance, heat capture, and customer/community benefits
- •Elena highlights bipartisan organizing power and says local leaders can’t ignore the issue without electoral consequences
Federal acceleration vs. local control: what happens when projects move to federal land
Kara asks how communities can respond when federal policy speeds approvals and uses federal land. The panel discusses the limits of local zoning, the bipartisan nature of executive actions, and the expectation of intensified backlash if local oversight is bypassed.
- •Schor argues local conditions and zoning processes are key levers—until federal land bypasses them
- •Elena says executive actions to use federal land began before and have since accelerated, reducing local control
- •She cites an example of a national park land swap proposal for data centers and calls it unchallengeable locally
- •A panelist notes interest in expanded federal oversight and describes state-by-state political volatility
- •Speakers predict political backlash could reshape party loyalties in some places
What “responsible” data centers would look like: costs, flexibility, and less demand
Kara closes by asking each panelist to describe what it would take for data centers to be genuine local value-adds. Ideas include strict tariffs and demand response, no incentives, careful siting, and shifting the core question from meeting demand to reducing it—alongside skepticism about nuclear timelines and costs.
- •Logan argues fair tariffs, demand flexibility, and transparent rules could reduce costs and limit peak-grid impacts
- •He critiques nuclear as too slow and expensive for near-term needs, and points to wind/solar/batteries/geothermal as available options
- •Schor emphasizes water protection, community benefits, no incentives, paying for buildout, and industrial-appropriate siting
- •Elena reframes the goal as demanding less compute and avoiding coal/diesel/gas-dependent designs with heavy cooling needs
- •A panelist argues companies should build safety and infrastructure up front rather than racing for speed and profit
Publisher timestamps
- 0:00Introduction to data centers
- 7:15Why data centers are causing backlash
- 9:15The secrecy behind data center dealsRead this chapter
- 16:20Data center and waterRead this chapter
- 30:30Who should pay for AI?
- 35:15Data centers and the power grid
- 41:05What guardrails should exist?
- 56:20What responsible data centers look like
