Better Offline
Better Offline

Sep 30, 2026 · 16 min

Listen from 6:08

Listen at 6:08

Anthropic’s losses expose the economics of AI scale

Monologue: Anthropic Lost $8bn In 2025, Was A Worse Business Than OpenAI

Anthropic’s reported losses and high cost of revenue raise broader questions about whether frontier AI businesses can become financially sustainable.

3 key takeaways
  1. 1Anthropic reportedly lost roughly $8 billion in 2025 while spending $2.75 to generate each dollar of revenue.
  2. 2Ed Zitron argues that Anthropic’s economics look weaker than OpenAI’s, despite both competing to build frontier AI systems.
  3. 3The leaked S-1 increases pressure to release the complete filing so investors and counterparties can assess the company’s condition.

Don't miss

The episode crystallizes its argument around Anthropic reportedly spending $2.75 for every dollar of revenue while losing roughly $8 billion in 2025.

The brief

Ed Zitron changes course after a leaked Anthropic S-1 reveals financial details he argues deserve scrutiny, especially from investors and the company’s economic counterparties.

The disclosures reportedly show Anthropic lost roughly $8 billion in 2025 and spent $2.75 to generate each dollar of revenue, putting its business model under pressure.

Zitron compares Anthropic unfavorably with OpenAI, framing the gap between frontier AI ambition and financial performance as a question of business sustainability.

The episode’s central demand is straightforward: release Anthropic’s complete S-1 so outside stakeholders can evaluate the company with more than leaked figures.

What was said on this episode

10 statements · 10 negative

  1. Ed Zitronon AnthropicNegative5:45

    Anthropic was a worse business than OpenAI in 2025.

    “This, shockingly, means that Anthropic was a worse business than OpenAI in 2025”

    Listen at 5:45

  2. Ed Zitronon AnthropicNegative6:58

    Anthropic’s economics have deteriorated over the past two years.

    “Anthropic's economics have decayed for the last 2 years”

    Listen at 6:58

  3. Ed Zitronon AnthropicNegative8:27

    Anthropic cannot presently afford its $518 billion compute commitments.

    “I don't see how this company becomes one that can afford its $518 billion in compute commitments”

    Listen at 8:27

  4. Ed Zitronon AnthropicNegative8:38

    Anthropic will likely rely on junk-grade or high-yield debt if it goes public.

    “This company will, if allowed to go public, likely lean on the very same junk-grade or high-yield debt”

    Listen at 8:38

  5. Ed Zitronon AnthropicNegative8:47

    Anthropic may issue $50–$100 billion of such debt annually.

    “at volumes of somewhere between $50 billion and $100 billion a year”

    Listen at 8:47

  6. Ed Zitronon AnthropicNegative8:55

    Anthropic will not become technology’s future, Google, or Microsoft.

    “Anthropic is not the future of technology, nor is it the next Google, nor is it the next Microsoft”

    Listen at 8:55

  7. Ed Zitronon AnthropicNegative9:08

    Anthropic is a cloud software company with volatile products and economics.

    “Anthropic is a cloud software company with volatile products and economics”

    Listen at 9:08

  8. Ed Zitronon Dario AmadeiNegative9:27

    Dario Amodei is manipulative and deceptive.

    “Dario Amadei is a manipulative and deceptive person”

    Listen at 9:27

  9. Ed Zitronon AI financial reportingNegative10:08

    Those responsible for misleading AI financial reporting should face imprisonment.

    “Someone needs to go to jail for this”

    Listen at 10:08

  10. Ed Zitronon OpenAI and AnthropicNegative10:40

    OpenAI and Anthropic are economic disasters waiting to happen.

    “We must treat OpenAI and Anthropic as what they are— economic disasters waiting to happen.”

    Listen at 10:40

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.

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

Anthropic’s losses expose the economics of AI scale · PodLume