Why No One Talks Cournot, Hollywood vs. Seedance 2.0, Micron’s $200B Bet | Jon Caramanica, Haseeb Qureshi, Spenser Skates, Celine Halioua, Ankur Goyal, Reed Duchscher

February 17, 2026

AI Summary

5 min read

The episode centers on the economics of frontier AI development, where a handful of labs are locked in an investment-driven competition that echoes classic oligopoly models. Hosts and guests unpack how this dynamic shapes spending, margins, and long-term industry structure, while touching on parallel shifts in hardware supply chains and AI-driven media production.

AI Lab Competition and the Cournot Framework
A small number of labs—primarily OpenAI, Anthropic, and Google DeepMind—compete by choosing supply levels rather than cutting prices. Each anticipates rivals’ next training run or inference capacity expansion and responds by ordering more GPUs or building additional data centers. Training runs create depreciating assets whose value drops once a stronger model launches, while inference operates as a higher-margin manufacturing business with variable costs tied to GPUs and power. The result is sustained high willingness-to-pay for frontier access among developers and enterprises, even as labs remain deeply unprofitable overall because new training spend outpaces inference revenue.

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

What you'll learn

1 (00:51) **Cournot Equilibrium in AI Labs** - Deep dive into oligopoly dynamics among frontier labs, training vs inference margins, and supply-side competition
2 (15:48) **Dwarkesh/Dario Clip on AI Economics** - Excerpt explaining why small numbers of firms avoid zero-margin equilibrium despite heavy losses
3 (18:58) **Nash Equilibrium & Beautiful Mind Clip** - Quick detour into game theory parallels with AI race dynamics
4 (23:48) **Timeline Debate: AI Automation & Context** - Reactions to Dwarkesh on software engineering vs other white-collar jobs
5 (29:54) **Song of the Week: "Granny Got Hit by a Bazooka"** - Breakdown of the viral Miami XO track and its meme lifecycle
6 (33:06) **Adani $100B AI Infrastructure Bet** - India's largest data center commitment and geopolitical implications
7 (35:30) **Micron $200B Memory Expansion** - Boise and New York fab announcements amid supply crunch

+ Full timestamped outline available in the app

Show Notes

(00:52) - Why is no one talking about the Cournot Equilibrium?
(23:53) - 𝕏 Timeline Reactions
(33:04) - Adani Group Unveils $100B AI Investment
(35:31) - Micron Bets $200B to End AI Bottleneck
(41:01) - Warner Bros Reopens Paramount Talks
(43:05) - 𝕏 Timeline Reactions
(01:00:54) - Thrive Capital Raises $10B in New Funding
(01:02:54) - 𝕏 Timeline Reactions
(01:15:20) - Hollywood Responds to Seedance 2.0
(01:29:08) - Jon Caramanica, born in 1975, is an American journalist and pop music critic for The New York Times, known for his insightful coverage of hip-hop and popular music. He discusses the viral ascent of "My Granny Got Hit with a Bazooka" by Miami XO, highlighting its rapid memeification and the artist's sudden rise from obscurity. Caramanica examines the song's adaptability across various formats, the music industry's response to such unexpected hits, and the evolving dynamics between virality and artist development.
(02:02:17) - Spenser Skates, co-founder and CEO of Amplitude, transitioned from algorithmic trading at DRW Trading Group to leading a product analytics company that went public in 2021. He discusses the benefits of taking Amplitude public, including enhanced talent acquisition and long-term stability, and emphasizes the importance of rapid innovation in the SaaS industry to maintain a competitive edge.
(02:24:20) - 𝕏 Timeline Reactions
(02:29:16) - Haseeb Qureshi, Managing Partner at Dragonfly Capital, a $4 billion crypto venture firm, discusses the firm's recent announcement of their fourth fund and the evolving landscape of crypto investments. He highlights the increasing intersection of crypto with AI, noting investments in platforms like Polymarket and the acquisition of OpenClaw by OpenAI. Qureshi emphasizes the enduring nature of crypto, its expansion into various sectors, and the growing adoption of stablecoins, particularly in countries with high inflation, despite regulatory challenges.
(02:42:42) - Celine Halioua, founder and CEO of Loyal, a biotech company developing drugs to extend dogs' lifespans, discusses the company's recent $100 million Series C funding round, bringing total funding to over $250 million. She emphasizes the primary goal of obtaining FDA approval for their first longevity drug, highlighting that much of the development process involves time-consuming steps like aging studies and regulatory responses. Halioua also notes the impact of GLP-1 drugs on public perception of aging biology, making the concept of aging interventions more tangible and understandable.
(02:51:33) - Ankur Goyal, founder and CEO of Braintrust, a company specializing in AI product observability, discusses the challenges of predicting AI product behavior post-launch and emphasizes the importance of monitoring re