Zitron Squeeze

A GPU can be sold to a customer in one country and used in another. That gap sits at the center of a much bigger argument about NVIDIA’s growth and export controls.

• The Megaspeed questions
• Enforcement and access
• Revenue versus destination
• Who funds demand?
• Access, accountability, and evidence

Source:

• Ed Zitron's Where's Your Ed At — Jensen's Bizarre Adventure

This podcast was created with Podkey. Make your own at https://podkey.fm

An AI company can have tens of billions in cash, fast-growing revenue, and still face a financing problem. The question is how long its spending commitments extend beyond the money coming in.

• OpenAI’s fundraising
• Anthropic’s commitments and margins
• The infrastructure financing test
• Following the revenue
• What users eventually pay

Source:

• Ed Zitron's Where's Your Ed At — Running Dry

This podcast was created with Podkey. Make your own at https://podkey.fm

Anthropic’s revenue is soaring. Its compute costs are rising faster. That gap is a useful place to start asking who’s earning a return on the AI boom, and who’s financing the wait.

• Where are the returns?
• Anthropic’s moving financial picture
• The compute bill
• How durable is demand?
• What customers actually pay
• Financing and exit incentives

Source:

• Ed Zitron's Where's Your Ed At — Exit Liquidity

This podcast was created with Podkey. Make your own at https://podkey.fm

The AI trade has a striking amount riding on a handful of companies. The question is whether the revenue beneath all that spending can catch up.

• A concentrated bet
• Where the returns have to come from
• Racks before power
• Debt against unfinished facilities
• Who pays whom
• Who bears the downside
• The missing receipts

Source:

• Ed Zitron's Where's Your Ed At — AI Had Better Work Out

This podcast was created with Podkey. Make your own at https://podkey.fm

A company can announce a $48 billion annualized revenue run rate in May and report $15.5 billion in revenue for the entire first half of the year. Both figures can be accurate. The trouble starts when we treat them as though they answer the same question.

• What a run rate measures
• The customers behind lab revenue
• What compute earns
• The stakes of the buildout

Source:

• Ed Zitron's Where's Your Ed At — Annualized Enron Rate

This podcast was created with Podkey. Make your own at https://podkey.fm

An AI company can sign a contract for compute years into the future. Someone still has to borrow the money to build the data center, buy the chips, and get the power connected.

• The price of money
• When a project needs a lender
• The labs’ financing wall
• Who is buying the compute?
• Risk travels through contracts
• The customer cash-flow test

Source:

• Ed Zitron's Where's Your Ed At — Credit Crunch

This podcast was created with Podkey. Make your own at https://podkey.fm

A GPU can be sold, financed, and counted as AI growth before it ever runs a customer’s workload. How much of this boom survives the trip from purchase order to paid invoice?

• Sold chips, usable capacity
• Revenue moving in circles
• Borrowing into the buildout
• The contract underneath the data center
• Customers funded by customers
• The labs at the center

Source:

• Ed Zitron's Where's Your Ed At — Dead Money

This podcast was created with Podkey. Make your own at https://podkey.fm

A gigawatt sounds concrete. So does a GPU sale. Yet neither phrase tells us whether a customer can use that compute today, or whether it is sitting behind an unfinished building, a missing power connection, and a financing promise.

• What does capacity actually mean?
• The pile of bought, uninstalled hardware
• Depreciation postpones the bill
• Who is actually buying compute?
• The overbuild question
• GPU demand versus useful demand

Source:

• Ed Zitron's Where's Your Ed At — Where're All The AI Chips?

This podcast was created with Podkey. Make your own at https://podkey.fm

A model gets pointed at a cybersecurity benchmark, finds a path through somebody else’s infrastructure, and suddenly the story becomes “the AI escaped.” That phrase can obscure the most important question: who built the setup, authorized the access, and watched the system while it ran?

• The Hugging Face incident
• Responsibility stays with the deployer
• Safety rhetoric and present harms
• The slowdown problem
• Hype meets public panic
• Legal accountability

Source:

• Ed Zitron's Where's Your Ed At — AI Is Already In Dangerous Hands

This podcast was created with Podkey. Make your own at https://podkey.fm

“AGI has arrived” makes a great headline. A trillion dollars of compute commitments makes a much harder spreadsheet. The question hanging over this boom is simple: who ultimately pays for all this capacity, and with what cash flow?

• AGI as a moving target
• Revenue concentration behind the boom
• The compute reset wall
• Circularity and the backlog story
• Where end-user demand stands
• What would break the narrative

Source:

• Ed Zitron's Where's Your Ed At — Concentration Risk

This podcast was created with Podkey. Make your own at https://podkey.fm