More
A video worth knowing

Jon Stewart presses Ed Zitron on AI's debt-funded buildout

Zitron argues that borrowing against chips and betting on future computing demand leave the boom exposed to more than disappointing chatbot sales.

Why this story? 892,846 views so far, and here is what is inside it.

Data Centers, Debt & the Risks of AI’s Finances w/ Ed Zitron | The Weekly Show with Jon StewartData Centers, Debt & the Risks of AI’s Finances w/ Ed Zitron | The Weekly Show with Jon Stewart

Things we can learn from the video

Press to jump

  1. 10:42Zitron explains the two main revenue routes for AI labs: subscriptions and metered access for developers.
  2. 13:54Data centers require enormous electricity supplies; Zitron uses a planned Abilene facility to illustrate the scale.
  3. 22:11Zitron reads Anthropic's filing as evidence of subsidized costs and substantial, uncancelable lease commitments.
  4. 24:06He describes debt secured against Nvidia chips and computing capacity leased back to AI labs.
  5. 37:37Stewart and Zitron consider whether interconnected borrowing could spread losses beyond individual AI companies.
  6. 49:57Zitron identifies chip and high-bandwidth memory supply constraints as further risks to the buildout.

"This is the worst episode of Shark Tank ever," Jon Stewart says as explains the financing behind AI's enormous infrastructure buildout.

The problem Zitron describes isn't simply that chatbots cost money to run. It's that companies are borrowing to build computing capacity on the assumption that AI businesses will keep needing, and paying for, much more of it.

In their Weekly Show conversation, Zitron describes arrangements involving and in which graphics processors serve as collateral for debt used to finance computing infrastructure, with capacity then leased to AI labs. His concern is that the apparent expansion rests on interconnected commitments rather than clearly demonstrated, profitable demand.

's public-offering filing gives that argument a concrete focus. Zitron points to long-term lease obligations and what he describes as subsidized operating costs. He also argues that venture-backed AI startups account for much of the labs' revenue, raising a further question: how durable are those sales if the customers themselves depend on investors' money?

Stewart pushes the discussion toward the consequences beyond Silicon Valley. Together they consider parallels with the dot-com bubble and the mortgage crisis, asking how trouble could spread through leveraged deals. Those are comparisons they explore, not evidence that the same outcome is inevitable.

The interview, which includes paid advertising breaks, has reached roughly 893,000 views. Amid the financing explanations, Zitron briefly picks up his cat, Hal, then returns to explaining the tiny units of text that AI companies charge customers to process.

Published on Trari

Where next?

What viewers say

  • New character unlocked in the Zitronverse
    @THEdavehogueSee it on YouTube
  • Said literally out loud in my car to myself: is that Ed Zitron with Joh Stewart?
    @fieldroaster21See it on YouTube
  • Ed and Jon finally unite… it’s a great day to be alive.
    @TheTaxManComethCPASee it on YouTube
2 more comments

Public comments from the video's page on YouTube

Or follow us on Telegram

The facts
Subscribers
902,000 ▲ 11,000 this week
Channel views, last 7 days
4,990,897
Channel views, pace per day
1,591,281 (on a normal day: 482,339)
Video published
October 28, 2026
Video views so far
892,846 (3 days online)
Video views, pace per day
237,468 (on a normal day: 420,125)

The numbers

The full interview's viewing figures sit below, alongside the channel's recent activity.

subscribers
902,000
▲ 11,000 in the last 7 days
total views
151,324,230
views per day
1,591,281
typical day: 482,339
the video
views
892,846
likes
16,865
comments
2,832
days online
3
subscribers: this channel and similar channels
Avi Yemini
943,000
Owen Jones
891,000
As of October 31, 2026 · Data source: Trari Social News, Trari Traffic Trends and Ranking Platform, Google YouTube
Subscriber, view, like and comment counts are YouTube's. Rates, usual days, rankings, comparisons and categories are Trari's own calculations.
Developed with YouTube

💡 Test your knowledge

What is the name of Ed Zitron's cat, who makes an appearance during the interview?