Price:

AI & TECH

AI energy surge forces fossil fuel revival

Wednesday, May 20, 2026 · from 5 podcasts
Simon Dixon Hard TalkTucker Carlson ShowWhat Bitcoin DidThis Week in AIThe a16z Show
  • AI’s power hunger is reversing 15 years of climate policy, reviving fossil fuels.
  • Bitcoin miners monetize stranded energy while waiting for AI infrastructure.
  • The grid’s mechanical core can’t handle digital demand, risking blackouts.

AI’s explosive growth is no longer just a tech story - it’s a national power crisis. Data centers now demand electricity at a scale that outstrips entire states, forcing a reckoning with the U.S. energy grid’s limits. In Utah, a single proposed facility would draw nine gigawatts - more than double the state’s current consumption - and 36 times the power of Boeing’s Everett plant, all while sprawling over 400 times the acreage.

This surge has shattered climate orthodoxy. Just 18 months after declaring net-zero mandates, financiers and officials now demand a fossil fuel revival. As Tucker Carlson notes, the same elites who blocked pipelines now push for LNG expansion, calling it a national security imperative. Kevin O’Leary frames the AI buildout as an arms race: China is constructing 400 gigawatts of coal power dedicated to AI, and the U.S. must match it or lose.

Bitcoin miners, meanwhile, are adapting to the new landscape. Harry Sudock explains that while AI needs fiber-connected urban campuses, Bitcoin thrives at the grid’s edge, using Starlink and stranded power. CleanSpark secures land and power contracts, runs miners to monetize electricity immediately, and waits for fiber to catch up - a strategy Sudock calls "squatting on power" until AI infrastructure matures.

The grid itself is the weak link. Drew Baglino of Heron Power calls it a pre-WWII mechanical system in a digital age - overbuilt, fragile, and blind. While batteries and chips evolved, transformers still rely on oil and copper. His solution: silicon carbide semiconductors to create a responsive, digital grid. But without a federal trust fund for energy infrastructure, the U.S. remains a patchwork of archaic systems incapable of supporting AI’s 100x compute demands.

"Bitcoin mining is the cockroach of compute - it thrives where others can’t go."

- Harry Sudock, What Bitcoin Did

The geopolitical stakes are rising. The Shanghai summit between U.S. tech elites and the CCP wasn’t about Taiwan or fentanyl - it was about securing the AI and robotics buildout. Simon Dixon argues the U.S. is in managed decline, with bond yields above 5% and foreign nations dumping Treasuries for gold. The dollar’s weakening allows massive money printing to bail out the financial system - all justified by the AI arms race.

Inside this shift, a new class is emerging. Nick Harris predicts the 1% will build $10 million private data centers, creating a "superhuman" productivity gap. Philip Johnston of StarCloud is already building orbital data centers in dawn-dusk orbit, using continuous solar power to bypass terrestrial bottlenecks. On Earth, workers face a purgatory: hired to train AI, then laid off when it reaches parity. Cloudflare cut 20% of its staff while reporting record revenue, a pattern repeating across the sector.

"We’re not building AI to serve people. We’re building people to serve AI."

- Tucker Carlson, The Tucker Carlson Show

The deeper crisis may be spiritual. Carlson warns that if machines absorb all intellectual labor, human purpose collapses. Young graduates now boo AI at commencement speeches - not out of fear of change, but of irrelevance. The promise of UBI or new jobs in robotics rings hollow. The real question isn’t whether the grid can handle AI - it’s whether society can.

#AI & Tech#Energy#Business#Regulation#Politics
CloudflareBitcoinChina

Source Intelligence

- Deep dive into what was said in the episodes

Simon Dixon Hard Talk
Simon Dixon Hard Talk

Silicon Valley Meets the CCP: What the Shanghai Summit Tells Us About the AI Arms Race • May 15

  • The Shanghai summit between US corporate elites and the CCP formalized economic ties to facilitate a managed transition to a multipolar world order, with the AI and robotics buildout as the central industrial project.
  • The White House announced China would resume buying US oil and LNG, which Dixon frames as a low-value, face-saving concession that merely returns trade to its pre-2025 status quo before China halted purchases.
  • Dixon claims Trump's trade policy led to record US small business bankruptcies, with 95% of tariff costs paid by them, while large multinationals offset costs via international production.
  • A key structural vulnerability Dixon identifies is the West's derivatives market for commodities, where paper contracts vastly exceed physical holdings, while China acquires physical gold without such overhang.
  • Dixon argues the AI arms race narrative justifies the enormous fiscal spending needed to bail out the financial system, with the AI data center buildout consuming vast energy and driving stock market concentration.
  • He frames the Iran conflict and Strait of Hormuz closure as a manufactured crisis to renegotiate global trade routes, force majeure contracts, and manage the transition away from a dollar-dominated system.
  • Dixon sees a fundamental asymmetry in US-China investment: Chinese investment in US companies grants the CCP covert voting influence, while US investment in China remains subordinate to strict CCP control and terms.
  • He identifies Palantir as a central actor building a privatized public-partnership technocratic state, testing its systems in Gaza, Ukraine, and Saudi Arabia, but notes it cannot control the CCP node.
  • He states the Clarity Act is the final legislative piece needed for the US surveillance state, with a current lobbying battle between finance and tech factions over who controls the resulting system.
Also from this episode: (3)

Markets (1)

  • Simon Dixon argues the US bond market shows severe stress, with the 30-year Treasury yield printing above 5% and the 10-year yield over 4.5%, which translates to 7% mortgage rates and a seized-up real estate market.

Inflation (1)

  • Dixon states US CPI registered at 3.8%, but he argues real inflation is far higher, citing a 77% increase in US beef prices since January as one example.

Protocol (1)

  • Dixon advises using hardware wallets for Bitcoin self-custody, storing 24-word seed phrases offline to maintain sovereignty outside the controlled ETF and banking system.
#AI Infrastructure#Big Tech#Macro#China#Energy
The Tucker Carlson Show
The Tucker Carlson Show

Tucker Carlson

DEBATE: Tucker vs Kevin O’Leary on the Dystopian AI Future Devouring American Energy and Jobs • May 14

  • Tucker Carlson argues that the closure of the Strait of Hormuz has created a severe global energy crisis, causing a net loss of 1.8 billion barrels of oil.
  • Carlson states that despite a 5% average rise in US homeowner energy costs, a powerful chorus from elected officials and financiers now demands a massive expansion of fossil fuel energy production to power AI.
  • Carlson cites a proposed Utah data center requiring 9 gigawatts of power, which he says is more than double Utah's total current energy consumption.
  • Carlson contrasts the Utah facility with the Boeing Everett plant, noting the data center would use 36 times the power while being over 400 times larger in acreage.
  • Carlson claims AI developers have failed to explain how the technology will improve average lives, instead framing it as an existential race against China.
  • Kevin O'Leary frames the Utah data center as a national security imperative, arguing the nation with superior AI compute power will win future wars and dominate the economy.
  • O'Leary states his data center will be energy independent, using low-cost stranded natural gas from the Ruby pipeline and new air-cooled turbines to avoid raising local electricity costs.
  • O'Leary defends tax incentives for large-scale projects as standard competitive practice among states to attract investment and jobs.
  • O'Leary argues AI will create millions of new high-paying jobs in fields like advanced robotics, medical science, and defense, countering predictions of mass job displacement.
  • Tucker Carlson counters that technological revolutions like the Industrial Revolution caused massive social disruption and world wars, and AI's potential to eliminate human purpose is a profound threat.
  • Carlson raises concerns that AI's capacity for deception, alignment problems cited by pioneers like Geoffrey Hinton, and its use for state surveillance represent more immediate dangers than a sci-fi takeover.
Also from this episode: (1)

AI Infrastructure (1)

  • O'Leary projects the first phase will cost $15 billion and create 10,000 construction and 2,000 maintenance jobs, financed by investors, not taxpayers.
#AI Infrastructure#Energy#China#Labor#Regulation
What Bitcoin Did
What Bitcoin Did

Danny Knowles

The Bitcoin Treasury Machine | Harry Sudock & Rory Murray • May 13

  • Harry Sudock says Bitcoin mining and AI differ in their energy stories: AI addresses insufficient power generation, while Bitcoin mining tackles inefficient power consumption. Both increase electron utilization but have distinct operational profiles.
  • Rory Murray argues AI's rise will decentralize Bitcoin's hash rate. Large energy-backed compute will prioritize AI for higher enterprise value, pushing Bitcoin mining to geographic and jurisdictional frontiers, creating a hub-and-spoke model.
  • CleanSpark's AI strategy involves greenfield development adjacent to mining sites, not retrofitting. Success requires four steps: power/land acquisition, leasing agreements, capital-intensive financing, and securing investment-grade tenants.
Also from this episode: (7)

Protocol (6)

  • CleanSpark's Bitcoin treasury management operates a dual strategy. Its 'spot plus' program enhances returns from monthly spot sales, while its 'yield program' aims to generate durable yield from its hodl by leveraging derivatives market volatility.
  • CleanSpark generates 500-600 Bitcoin monthly from mining. A portion is sold for OPEX and CAPEX, while the team deploys strategies like selling short-dated covered calls to extract additional margin from the Bitcoin before conversion.
  • Rory Murray says their covered call strategy is self-reinforcing because they have an operating business that prints Bitcoin. If calls are exercised during a parabolic move, they can pause spot sales for months, replacing the called-away Bitcoin with future production.
  • Rory Murray states Bitcoin's liquidity and 24/7 trading make it superior collateral for loans. He says institutional Bitcoin-backed loan rates have compressed from 9-11% to around 6%, citing CleanSpark's recent paper at 'software plus 3.55%.'
  • The pair believe Bitcoin should trade at a lower loan rate than corporate credit due to its over-collateralization, automatic liquidation, and 24/7 global liquidity, which creates a near-seamless, lossless collateral liquidation mechanism.
  • Rory Murray outlines a treasury flywheel: use appreciating Bitcoin to borrow depreciating dollars, deploy dollars into appreciating assets like AI data centers, and use the revenue to fuel further growth and Bitcoin acquisition.

AI Infrastructure (1)

  • Harry Sudock says CleanSpark's digital asset management is not an internal hedge fund. It is designed to feed and enhance mining profitability, fund expansion, and maximize the huddle's potential by taking risks hedged by the operating business.
#Macro#BTC Markets#AI Infrastructure#Energy
This Week in AI
This Week in AI

How the 1% Will Own Compute (and What It Means for You) • May 13

  • Nick Harris argues AI interaction models will demand 100x more compute and energy than current systems, creating a bottleneck his photonic computing chips and Philip Johnston’s space solar data centers aim to solve.
  • Philip Johnston explains StarCloud’s orbital strategy uses a dawn-dusk sun-synchronous orbit for continuous solar power, drastically reducing battery needs compared to terrestrial solar projects.
  • Thinking Machines' new 'Interaction Model' processes audio, video, and text in real-time micro-turns, decoupling interaction from background thinking to enable continuous multimodal context.
  • Anastasios Angelopoulos argues China's open-source AI models lag US proprietary labs by roughly two quarters, but this gap could become existential if frontier model improvements plateau.
  • Nick Harris predicts compute polarization where the 1% will own personal $10 million data centers, enabling superhuman productivity while the broader power grid cannot support democratized 100x scaling.
  • Jason Calacanis observes a tenfold productivity gap between AI-first and AI-averse employees within his venture, forcing systematic retraining to prevent obsolescence.
  • Nick Harris and Jason Calacanis forecast an entrepreneurship boom as AI-driven layoffs push talent toward small, autonomous teams that can profitably operate outside traditional corporate structures.
  • Philip Johnston states current text-to-CAD models handle simple components like screws but fail at complex tasks like designing a full 200 kW satellite with deployable radiators.
  • Anastasios Angelopoulos frames the core societal challenge as decoupling labor from value creation, requiring careful incentive redesign to transition to an abundance economy without destabilizing collapse.
  • Jason Calacanis cites Cloudflare cutting 20% of its workforce while reporting record revenue as emblematic of the concurrent rise of superintelligent models and social unrest.
Also from this episode: (1)

AI Infrastructure (1)

  • Philip Johnston details hardware modifications for space compute: stripping casings and heatsinks, radiation shielding, and ruggedizing chips for launch vibration.
#AI Infrastructure#Chips#Labor#Enterprise#Models
The a16z Show
The a16z Show

Energy, Minerals, and the Physical Stack Behind AI • May 13

  • Turner Caldwell of Mariana Minerals says the U.S. is 50 years behind China on critical minerals supply, a gap that persists even if permitting and finance accelerate.
  • Drew Baglino of Heron Power says grid infrastructure relies on pre-World War II mechanical systems, creating a fragile, overbuilt network with overseas suppliers, while innovation has only occurred at the grid's edge.
  • Caldwell suggests applying the regulatory and incentive toolkit used for the oil and gas industry over the last 50 years to a new minerals mandate, to mobilize private capital with long-term market confidence.
  • Baglino advocates for durable industrial policy, federal-state identification of energy/manufacturing zones for co-located supply chains, and a federal highway trust fund model for grid infrastructure.
Also from this episode: (7)

AI Infrastructure (1)

  • Mariana Minerals develops three software systems to automate mining and refining: Capital Project OS for project lifecycle, Plant OS for refinery control via reinforcement learning, and Mine OS for autonomous mining operations.

Startups (2)

  • Caldwell argues vertical integration from mining to refining avoids market inefficiencies, and that software adoption in these industries hinges on embedding engineers directly with operating teams to control culture and tool design.
  • Caldwell and Baglino cite the Tesla model: a belief in innovating archaic systems, appetite for risk enabling fast decisions, and a firm commitment to fighting through challenges for worthy outcomes.

Chips (2)

  • Heron Power builds solid-state transformers using silicon carbide semiconductors to replace steel, oil, and copper in grid power conversion, targeting data centers and large-scale energy installations.
  • Baglino says the U.S. developed silicon carbide technology and should commercialize it domestically; losing that manufacturing means ceding all downstream benefits to other countries.

Labor (2)

  • Baglino contends labor cost differentials between U.S. and Chinese factories are under 10% of COGS, with competitiveness driven by supply chain co-location, not wages.
  • Baglino notes that building a U.S. industrial workforce requires hiring from analog industries like high-speed bottling or syringe manufacturing, not existing power electronics talent pools.
#AI Infrastructure#Startups#Chips#Energy#Regulation

Related Stories