Recording intelligenceAI-generated brief · check the source for context

What Happens When the AI Boom Runs Out of Money

1:25:54 recording · EN · 2 speakers

Watch the original

Executive Summary AI
  • Ben argues it would be problematic for the US to win the AI race outright, because a meaningful military superiority makes China's game-theory optimal move blowing up TSMC, and US dependence on Chinese manufacturing is too deep to unwind outside a conflict.
  • He likes the current equilibrium instead: OpenAI and Anthropic on the frontier, Google uncertain, Grok and Meta chasing, and capable Chinese labs distilling to stay about six to nine months behind.
  • His real worry is timing, not capability — the industry has worked down the capital curve from free cash flow to the debt markets to Google issuing equity and a $500 billion vehicle tapping pension funds and insurance floats, and if revenue does not arrive before the money runs out there is a blow-up.
  • On supply, he says risk never disappears but moves: TSMC's conservatism after its 2020-22 expansion pushed risk onto big tech as foregone revenue, and that scarcity is what may finally get Intel and Samsung logic a first major customer.
  • He calls NVIDIA's maintained margins unnatural — its equity and offtake deals with neoclouds are a disguised price cut — and says its real competitors are Google and Amazon with cheaper capital, while abundant US power has bought them time and power is the lasting asset this build-out leaves behind.

Brief overview

American AI dominance would be destabilizing; the real risks are commodity economics, funding timing and compute dependency.

  1. Risk never disappears, it only movesTSMC refused the overcapacity risk of 30-year fabs, so big tech carries it as foregone revenue and profits.
  2. Consumers will not pay, so advertising winsBen says OpenAI replayed the Dropbox story at 100x and should have leaned into ads as soon as ChatGPT hit.
  3. Compute and intelligence are commodity businessesIn commodity markets price is set by the marginal supplier and cost to serve is all that matters — like ships, memory and bandwidth.

Questions this recording answers

10 questions, each answered where it is said
Why would a decisive U.S. victory in the AI race be dangerous?

Very problematic, he argues. Take the fantastical case: controlling AI makes your military better than anyone else's. China's game-theory optimal response is then simply to blow up TSMC. If the U.S. reaches meaningful superiority, particularly from a military and national security perspective, that is very dangerous for the world.

Could the U.S. escape its dependence on Chinese manufacturing?

Not outside a conflict. Talk of fixing fabs, actuators and other precursors is magical thinking; sourcing from the U.S. while your competitor sources from China is such a disadvantage you simply won't do it. Apple was browbeaten into some iPhone production in India, but it is diversifying, not truly leaving China.

How do the U.S. and Chinese AI labs compare right now?

OpenAI and Anthropic are clearly on the frontier, who knows what is happening with Google, and Grok and Meta are chasing. The Chinese are very capable and very smart, distilling these models to stay about six to nine months behind. He calls that a pretty good equilibrium, generally favorable to the U.S.

Are open-source models really free to use?

No. Calling them free confuses R&D with marginal cost. You still have to run inference, like GLM or Kimmy, and Kimmy is very expensive to serve, with a significantly higher cost per answer. You are not paying the R&D that created the AI, but you definitely pay the inference.

What does Berkshire's move from See's Candies to BNSF say about Google?

See's is a tremendously high margin business, but the absolute profit is capped with no reinvestment runway. BNSF's margins are far worse, yet in one year it threw off more free cash than See's had over its entire lifetime. Google search may be See's Candies, and AI the BNSF.

Does AI's strength at coding and math mean it generalizes?

He is not fully convinced. AI is clearly incredible at coding and obviously very good at math, but those are verifiable domains. The compelling evidence that this translates cleanly to unverifiable domains, or ones with very long verification loops, is missing. Chess and Go were knowable, bounded domains solved by scale.

What is aggregation theory in a nutshell?

Zero marginal costs mean you already have distribution; your problem is demand, which is why you pay for ads. In a world of abundance the hard problem is discovery, so the companies that solve discovery in their domain dominate it and get a virtuous feedback loop. Transaction costs are zero too.

Why has no great business model emerged for casual chatbot users?

Frustrating but not surprising. Silicon Valley relearns two things every ten years: consumers do not want to pay for software, and consumers do not care about being productive. This market should be supported by advertising. OpenAI replayed the Dropbox story at a hundred times scale, selling many subscriptions but not enough.

Why might the compute shortage end up rescuing Intel?

Because TSMC did not invest over the last few years, the shortages will be so acute that big tech companies foregoing enormous revenue and profits will go through the pain of getting Intel, and Samsung's logic, up to speed. Scarcity ultimately saved Intel; he expects a major partner announced this year.

Is NVIDIA actually maintaining its margins in 2026?

Not really. The circular financing — a 25% backstop, equity in the neoclouds, guaranteeing to buy their compute through 2030 — exists so those entities get a lower cost of capital, which means NVIDIA assumed the risk. Taking on risk has a price, so holistically that is a price cut.

Key Quote

“To blow up TSMC.”

— Ben, What Happens When the AI Boom Runs Out of Money, at 0:17▶ watch at 0:17
Key Quote

“I think we need to push forward because we can't go back.”

— Ben, What Happens When the AI Boom Runs Out of Money, at 18:55▶ watch at 18:55
Key Quote

“Number one, consumers do not want to pay for software.”

— Ben, What Happens When the AI Boom Runs Out of Money, at 26:53▶ watch at 26:53
Key Quote

“NVIDIA's position is, I think, definitely unnatural.”

— Ben, What Happens When the AI Boom Runs Out of Money, at 1:15:41▶ watch at 1:15:41

Cite this

“What Happens When the AI Boom Runs Out of Money.” https://www.youtube.com/watch?v=h-0NZ-oIjlk. Transcript summary by WhipScribe, https://whipscribe.com/transcript-pages/what-happens-when-the-ai-boom-runs-out-of-money.html.

Report this page
What is wrong with this page?