Better Offline
Better Offline

Sep 9, 2026 · 22 min

AI’s losses test the limits of crisis management

Managing The Situation

The episode asks whether familiar startup comparisons conceal an AI business model that requires unusually large and potentially unsustainable capital.

3 key takeaways
  1. 1AI’s enormous losses are often defended through comparisons with early Uber and Amazon Web Services.
  2. 2Those comparisons may obscure differences in AI’s business model, scale, and capital requirements.
  3. 3The broader argument is that institutions increasingly manage crises through reassuring narratives rather than confronting complexity.

Don't miss

Zitron challenges the idea that AI’s losses can be explained away by comparing them with early Uber and Amazon Web Services.

The brief

Ed Zitron opens with a monologue after a guest canceled, using the format to examine how institutions and industries explain crises without resolving their underlying consequences.

The episode’s central target is the defense of AI’s enormous losses, especially arguments that compare today’s spending with early Uber and Amazon Web Services.

Zitron questions whether those comparisons hold: AI may not simply be an early-stage business absorbing losses, but a model demanding unprecedented capital without a clear path to sustainability.

The discussion places AI’s financial story within a wider pattern of crisis management, where familiar narratives can make instability appear normal instead of forcing harder questions.

The episode leaves a pointed test for the industry: can AI demonstrate a durable business model, or will rhetorical reassurance continue substituting for evidence?

What was said on this episode

28 statements · 25 negative · 1 mixed · 2 neutral

  1. Ed Zitronon Unresolved societal problemsNegative3:29

    Unresolved problems compound rather than disappear.

    “The problems actually compound.”

    Listen at 3:29

  2. Ed Zitronon Managed systemic problemsNegative3:29

    Managed problems compound rather than disappear.

    “the problems actually compound”

    Listen at 3:29

  3. Ed Zitronon Government crisis fundingNegative4:26

    Government funds preserved the financial system rather than fixing underlying social problems.

    “Government funds were used to manage a systemic crisis by making sure the system survived, rather than building a better society or fixing the system”

    Listen at 4:26

  4. The Great Financial Crisis was managed through zero rates and bailouts.

    “The Great Financial Crisis was managed by dropping interest rates to zero and bailing out banks and automakers”

    Listen at 5:34

  5. COVID policy used the loosest possible lockdown protocols.

    “COVID was managed by the loosest lockdown protocols possible”

    Listen at 5:49

  6. Ed Zitronon Financial system liquidityNegative6:43

    Easy money saturated the financial system in 2021.

    “The net result was that in 2021, the financial system got saturated with easy money.”

    Listen at 6:43

  7. Ed Zitronon Corporate price increasesNegative6:53

    Corporations used inflation concerns to raise prices broadly.

    “corporations have been using supply chain crises in the specter of inflation to raise prices across the board”

    Listen at 6:53

  8. Ed Zitronon Corporate price increasesNegative7:00

    Corporate greed, rather than inflation alone, drove price increases.

    “the real reason was sheer, unabated greed”

    Listen at 7:00

  9. Ed Zitronon Federal Reserve interest-rate increasesNegative7:17

    Federal Reserve rate increases led to major technology-sector job losses.

    “By the end of 2022, the Federal Reserve would raise interest rates by a dramatic 4.25%, leading to tens of thousands of people losing their jobs in the tech sector alone.”

    Listen at 7:17

  10. Ed Zitronon Corporate price increasesNegative7:38

    Corporate price increases mostly drove the rise in consumer prices.

    “the increase in pricing was mostly driven by, get this, corporations raising prices”

    Listen at 7:38

  11. Ed Zitronon OpenAI and Anthropic lossesNegative12:49

    OpenAI and Anthropic lose tens of billions of dollars annually.

    “OpenAI and Anthropic are losing tens of billions of dollars a year”

    Listen at 12:49

  12. Ed Zitronon AWSNeutral13:01

    AWS spent $29.7 billion in capex before becoming profitable.

    “Amazon Web Services' case, it was $29.7 billion of total capex between the year it started in 2003 and the year it became profitable in 2015.”

    Listen at 13:01

  13. Ed Zitronon Generative AI hallucinationsNegative13:16

    Generative AI hallucinations make LLMs inherently untrustworthy.

    “generative AI's hallucinations make massive mistakes that make LLMs inherently untrustworthy”

    Listen at 13:16

  14. Current artificial intelligence is simultaneously limited and highly capable.

    “This artificial intelligence is actually kind of stupid, but it's also super smart.”

    Listen at 13:33

  15. Ed Zitronon AI companiesNegative14:48

    AI companies do not have a secret overarching plan.

    “They don't, just want to be clear that they don't”

    Listen at 14:48

  16. Ed Zitronon Systemic failuresNegative15:00

    Systemic failures result from lacking genuine planning and forethought.

    “Every systemic failure has come from a lack of any real plan, strategy or forethought.”

    Listen at 15:00

  17. Ed Zitronon AI industryNegative15:33

    AI is repeating the systemic failures preceding the financial crisis.

    “We are watching the very same thing happen again with AI.”

    Listen at 15:33

  18. Ed Zitronon OpenAI and Anthropic IPO financingNegative16:11

    Debt-financed public listings will eventually leave OpenAI and Anthropic without money.

    “This is a deeply dangerous way of managing two companies that will eventually run out of money.”

    Listen at 16:11

  19. Ed Zitronon Anthropic and OpenAI compute commitmentsNegative16:17

    Anthropic and OpenAI will need hundreds of billions annually for compute commitments.

    “Anthropic and OpenAI will need hundreds of billions of dollars a year to meet their $1.3 trillion of compute commitments.”

    Listen at 16:17

  20. Ed Zitronon OpenAI and Anthropic debtNegative16:51

    OpenAI and Anthropic likely cannot service required debt without profitability.

    “unless they magically find ways to become profitable and generate massive amounts of cash, I don't see how they service the debt they're going to need.”

    Listen at 16:51

  21. Ed Zitronon OpenAI and Anthropic growthNegative17:52

    OpenAI and Anthropic are experiencing slowing growth.

    “Both companies are seeing growth decelerate”

    Listen at 17:52

  22. Ed Zitronon OpenAI and Anthropic enterprise revenue concentrationNegative17:54

    The top 1% of customers generate 80% of OpenAI and Anthropic enterprise revenue.

    “80% of their enterprise revenues per ramp, as I discussed last week, come from the top 1% of their customers”

    Listen at 17:54

  23. Ed Zitronon OpenAI and Anthropic investment-grade ratingsNegative18:24

    Investment-grade ratings would only delay OpenAI and Anthropic running out of money.

    “Any attempts to mark OpenAI or Anthropica's investment grade will simply delay the inevitable point at which they run out of money.”

    Listen at 18:24

  24. Ed Zitronon OpenAI and Anthropic compute commitmentsNegative18:41

    OpenAI and Anthropic’s compute commitments require payment regardless of actual demand.

    “Said commitments are also take-or-pay agreements”

    Listen at 18:41

  25. Ed Zitronon Anthropic and OpenAI take-or-pay data-center contractsNegative18:44

    Take-or-pay contracts require Anthropic and OpenAI to pay providers regardless of demand.

    “regardless of whether they have the demand to actually need the data centers, Anthropica and OpenAI will have to pay Amazon, Google, Microsoft Cerebris, CoreWeave, or any number of other providers once the capacity comes online.”

    Listen at 18:44

  26. Ed Zitronon AI financing systemNegative19:03

    The AI system will appear healthy until a participant cannot raise funds or pay.

    “everything will seem like it's fine right up until somebody can't raise money and make a payment to either a Neo, Cloud, Hyperscaler, or AI lab.”

    Listen at 19:03

  27. Ed Zitronon AI bubbleNegative19:49

    The AI bubble depends on circular spending among a small number of companies.

    “the AI bubble is based on the whims of maybe a few hundred companies spending money on two companies to justify five companies spending money with one company”

    Listen at 19:49

  28. Ed Zitronon AI market correctionNegative20:45

    AI-industry failures will cause a correction larger than the dot-com bubble.

    “The failure of everyone here to meet this moment with clarity and purpose will lead to a market correction that dwarfs the dot-com bubble easily”

    Listen at 20:45

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.