Thinking On Paper

Burning Earth: How AI Helps Big Oil Drill More


Listen Later

Exxon and Chevron are using Microsoft AI to extract more oil. Faster. Cheaper. The goal: every last drop.


Holly and Will Alpine (Enabled Emissions) paint a stark picture with Microsoft's own numbers:


Exxon deal: 50,000 barrels/day = 6.4 million tonnes CO₂/year

Chevron deal: 400,000 barrels/day = 51 million tonnes CO₂/year


Microsoft's entire FY23 footprint: 17 million tonnes.

Carbon removal booked over 15 years: 5 million tonnes.


Those two deals alone dwarf both numbers.


In Saudi Arabia, Aramco's CEO says AI kept production costs at $3/barrel for two decades. AI makes fossil fuels competitive. It weakens clean energy economics.


AI touches every stage of fossil fuel: exploration, drilling, extraction, refining, distribution.


We talk about:

- How AI cuts production costs 10% while boosting reserves 5%

- Why US oil production tripled since 2007 (AI is a major factor)

- Projects that took 18 months now take 2 weeks

- How aging oil fields stay profitable decades longer

- Why Microsoft's "energy principles" are meaningless (they ignore Scope 3)

- What guardrails could look like (restrict exploration/extraction, allow safety/methane reduction)


Holly and Will left Microsoft to start Enabled Emissions. They're not asking companies to break contracts. They're asking for reasonable guardrails.


The lie you're fed: AI will solve climate change.

The truth: AI is accelerating fossil fuel extraction.


It's ugly. It's real. Learn more.


---


Guests: Holly & Will Alpine, Founders Enabled Emissions (Former Microsoft)

Topics: AI, climate, oil, emissions, Microsoft, Exxon, Chevron, Aramco, enabled emissions

--


Other ways to connect with us:

  • ⁠Listen to every podcast⁠

  • Follow us on ⁠Instagram⁠

  • Follow us on ⁠X⁠

  • Follow Mark on ⁠LinkedIn⁠

  • Follow Jeremy on ⁠LinkedIn⁠

  • Read our ⁠Substack⁠

  • Email: [email protected]


...more
View all episodesView all episodes
Download on the App Store

Thinking On PaperBy Mark Fielding and Jeremy Gilbertson