Better Offline
Better Offline

Sep 4, 2026 · 15 min

AI’s growth rests on a dangerously concentrated customer base

Monologue: Concentration Risk

The episode argues that AI’s apparent expansion may depend on a narrow, financially fragile network of customers, infrastructure providers, investors, and lenders.

3 key takeaways
  1. 1OpenAI and Anthropic reportedly draw roughly 80% of enterprise revenue from their top 1% of customers.
  2. 2Concentrated demand makes AI infrastructure growth vulnerable when a small number of companies or buyers pull back.
  3. 3The ecosystem’s dependence on cloud platforms, venture capital, and debt makes its stability harder to assess.

Don't miss

The episode’s sharpest claim is that roughly 80% of OpenAI and Anthropic’s enterprise revenue reportedly comes from their top 1% of customers.

The brief

Ed Zitron opens by reframing the AI boom as a concentration problem: too much of the ecosystem may depend on too few companies, technologies, and customers.

The episode highlights claims that roughly 80% of OpenAI and Anthropic’s enterprise revenue comes from their top 1% of customers, leaving growth exposed to a narrow demand base.

That concentration extends beyond customers to cloud infrastructure, venture capital, and debt, creating a chain in which weakness at one link could pressure the rest.

References to Google Cloud Platform and developments involving GoPro broaden the discussion from AI companies to the infrastructure and technology dependencies surrounding them.

The central takeaway is that AI’s visible scale may overstate its resilience: a large ecosystem can still rest on a small number of financially important relationships.

What was said on this episode

23 statements · 22 negative · 1 neutral

  1. Ed Zitronon concentration riskNegative2:50

    Concentration risk can severely harm businesses or portfolios when key exposures fail.

    “becoming overly reliant on a few investments, customers, or particular business lines to the point that without them, your business or portfolio would suffer massive harms or just explode”

    Listen at 2:50

  2. Ed Zitronon AI bubbleNegative3:25

    The AI bubble will unravel in coming months and years

    “the AI bubble unravels”

    Listen at 3:25

  3. Ed Zitronon OpenAI and Anthropic AI-startup revenueNegative4:37

    OpenAI and Anthropic revenue from AI startups depends on those startups raising venture capital.

    “the money made by Anthropica OpenAI from an AI startup in that 1% spend is contingent on their continued ability to raise venture capital dollars”

    Listen at 4:37

  4. Ed Zitronon Enterprise AI spendingNegative4:48

    Most enterprises spend relatively little on AI

    “the vast majority of enterprises, which is where the real money is in software and the real growth is, just don't spend that much money on AI”

    Listen at 4:48

  5. Ed Zitronon OpenAI and Anthropic enterprise revenueNegative5:16

    Most OpenAI and Anthropic revenue depends on hundreds of customers with outsized AI-token spending.

    “80% of OpenAI and Anthropics enterprise revenues, which make up the vast majority of their total revenues, are dependent on what are likely hundreds of customers spending outsized amounts of money on AI tokens”

    Listen at 5:16

  6. Ed Zitronon Non-technology companies' AI spendingNegative5:50

    Few non-technology companies are willing to spend heavily on AI.

    “outside of the tech and AI world, very few companies are willing to pay very much for AI”

    Listen at 5:50

  7. Ed Zitronon Technology industry AI spendingNegative6:13

    The technology industry is largely selling AI spending to itself.

    “the only people the tech industry can get to spend money on AI is the tech industry itself”

    Listen at 6:13

  8. Ed Zitronon OpenAI and Anthropic customersNegative6:22

    OpenAI and Anthropic customers may shift to cheaper open-source or on-device models

    “these customers are also prime targets to move to either cheaper models that they train themselves, because they're open source, or eventually on-device models”

    Listen at 6:22

  9. Ed Zitronon OpenAI and Anthropic revenue concentrationNegative6:42

    80% of OpenAI and Anthropic revenue relies on unstable spending and capital

    “80% of the revenue concentration depends on spending and capital that varies from unreliable to actively unstable”

    Listen at 6:42

  10. Ed Zitronon OpenAI and Anthropic compute-supplier exposureNegative6:50

    OpenAI and Anthropic create major concentration risk for their compute suppliers.

    “these two AI labs represent a massive concentration risk for Microsoft, Google, Amazon, Oracle, CoreWeave, and anyone else that sells compute to them”

    Listen at 6:50

  11. Ed Zitronon OpenAI and Anthropic compute commitmentsNegative6:57

    OpenAI and Anthropic have $1.1 trillion in compute commitments tied to concentrated demand.

    “Anthropic and OpenAI signing over $1.1 trillion worth of compute commitments based on demand that's mostly coming from a very small subset of customers”

    Listen at 6:57

  12. Ed Zitronon Anthropic and OpenAI compute agreementsNegative7:06

    The compute commitments appear to be take-or-pay agreements

    “These are, from what I can tell, take-or-pay agreements”

    Listen at 7:06

  13. Ed Zitronon OpenAI and Anthropic compute agreementsNegative7:06

    OpenAI and Anthropic’s compute agreements require payment regardless of usage or revenue.

    “These are, from what I can tell, take-or-pay agreements where they agree to buy that compute capacity, regardless of how much capacity they actually end up using and how much revenue they actually bring in”

    Listen at 7:06

  14. Ed Zitronon Anthropic and OpenAI profitabilityNegative7:18

    Anthropic and OpenAI lose tens of billions of dollars annually

    “both Anthropic and OpenAI are woefully unprofitable to lose tens of billions of dollars a year”

    Listen at 7:18

  15. Ed Zitronon Microsoft AI revenue exposure to OpenAINegative7:25

    OpenAI accounts for about 70% of Microsoft’s AI revenue.

    “OpenAI's compute spend and revenue share represent about 70% of Microsoft's AI revenues in fiscal year 26”

    Listen at 7:25

  16. Ed Zitronon Microsoft, Google, and Amazon revenue exposureNegative8:09

    OpenAI and Anthropic are expected to generate over $174 billion for Microsoft, Google, and Amazon in 2027.

    “Anthropic and OpenAI account for over $174 billion worth of expected revenues from Microsoft, Google, and Amazon in 2027”

    Listen at 8:09

  17. Ed Zitronon Anthropic and OpenAI compute commitmentsNegative8:37

    Compute-contract payment pressure will begin next year and worsen monthly

    “That's going to start happening next year and get dramatically worse month after month as capacity starts powering up and they start actually having to pay for it”

    Listen at 8:37

  18. Ed Zitronon OpenAI and Anthropic compute commitmentsNegative9:19

    OpenAI and Anthropic’s compute commitments resemble subprime mortgages.

    “OpenAI and Anthropic's massive compute commitments are the subprime mortgages of the AI bubble”

    Listen at 9:19

  19. Ed Zitronon NVIDIA GPU customersNegative10:15

    NVIDIA customers cannot fund GPU purchases solely through cash flow.

    “NVIDIA's customers are no longer able to buy its GPUs through cash flow alone”

    Listen at 10:15

  20. Ed Zitronon AI concentration risksNegative11:18

    The unfolding AI concentration risks will produce the warned-about outcome

    “I think it's going to be the thing I've been warning about”

    Listen at 11:18

  21. Ed Zitronon AI industryNegative11:28

    The AI industry represents a catastrophic misallocation of capital.

    “I think this is a catastrophic misallocation of capital”

    Listen at 11:28

  22. Ed Zitronon AI financial structureNegative11:40

    The AI financial structure will eventually collapse.

    “when this eventually falls apart”

    Listen at 11:40

  23. Ed Zitronon Jane Street–CoreWeave relationshipNeutral11:59

    Jane Street is a major CoreWeave customer and investor.

    “Jane Street's a major customer of CoreWeave and an investor in CoreWeave”

    Listen at 11:59

Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.

Weekly email

State of the Pod, every Monday

The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.

Listen to the full episode and explore every guest, topic, and moment on PodLume.