Synthesis
Intelligence as a commodity
The energy-and-price argument: if machine intelligence is abundant, it is priced toward the cost of the energy that produces it, so selling tokens is a low-margin, utility-like bu…
This is mostly Sam Lessin’s thesis, repeated across 16 episodes by one co-host of one show. His line: if intelligence is abundant it is a commodity priced to the cost of energy , so producing it is low-margin and whoever has cheap energy wins . He compares AI to power and cable but says it is worse, because those are regional or legally protected monopolies while AI is globally liquid and easy to transport . By his account the labs are ‘utilities with none of the financial benefits of being utilities’ , and the money ends up in what people build on top . He floats two exits, regulatory capture or ‘start up, cash in, bro down’ . Dave Morin’s version is practical: anyone other than OpenAI and Anthropic is an ‘arms dealer’ renting compute and models . Sam himself concedes nobody knows the marginal cost per token at which this bites and that valuations may rise on hope regardless . The only outside voice is Ben Broca’s figure that AI costs about a thousandth of a human salary .
Key points
- Sam argues abundant intelligence is priced toward the cost of energy and is a low-margin business — 5 episodes. (Sam · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 13:12)
- AI is a worse business than a utility or cable because it has no regulation or physical barriers and is globally liquid — 5 episodes. (Sam · Anthropic’s Bet on Coding Is W (Mar 2026) ▶ 24:02)
- Value accrues to what is built on top of the commodity input, not to the seller of it — 2 episodes. (Sam · Anthropic’s Bet on Coding Is W (Mar 2026) ▶ 28:35)
- Labs are described as utilities, or ‘arms dealers’, renting models and compute — 2 episodes. (Sam · Open Source AI Wins. Now Labs (Jul 2026) ▶ 09:15)
- Sam says he does not care which model he uses, which he takes as proof of commodity status — 2 episodes. (Sam · AI Frontier Lab Is Essentially (Jul 2026) ▶ 45:48)
Frameworks
- Two ways a commodity utility business responds: regulatory capture or the South Park answer — Sam says a competitive, globally shippable utility can either get regulatory capture (only approved models legal, like a state monopoly) or follow the South Park model: start up, cash in, bro down. (Sam · Elon Musk vs OpenAI Trial, Goo (May 2026) ▶ 26:31)
- Regulatory capture: only one safe model usable in a region
- South Park answer: start up, cash in, bro down
Stories & examples
- Jess reports that OpenAI CFO Sarah Friar told her OpenAI is looking at licensing IP for drug discovery, which Sam reads as an acknowledgement that base intelligence is a commodity. (Jess · OpenClaw vs Meta vs OpenAI: Th (Feb 2026) ▶ 51:48)
Numbers
- AI cost relative to a human salary: 1000 times cheaper (said ‘like 1,000th of the price’) (self-reported, approx.) (Ben Broca · Henrik Werdelin & Ben Broca (Mar 2026) ▶ 07:48)
Caveats & warnings
- Sam says it is possible intelligence is a cheap commodity, but there is a lot nobody knows, including the marginal cost per token at which it becomes a commodity. (Sam · OpenClaw vs Meta vs OpenAI: Th (Feb 2026) ▶ 47:52)
- Sam says it may not matter if everyone puts their money in hope and hope keeps going up. (Sam · OpenAI’s $852 Billion Valuatio (Apr 2026) ▶ 37:23)
- Sam says the AI valuations are narrative-driven hope, and none of it makes financial sense. (Sam · OpenAI’s $852 Billion Valuatio (Apr 2026) ▶ 37:16)
How much to trust this
- Evidence: 16 episodes, 2 independent speakers (the four hosts and unidentified host speech count as one voice, the show; guests counted individually).
Sources
- (01)
OpenClaw vs Meta vs OpenAI: The Personal Agent Wars Heat Up
45:46 into “OpenClaw vs Meta vs OpenAI: The Personal Agent Wars Heat Up”
- (02)
Open Source AI Wins. Now Labs Are Running to Washington |Thinking Machines, Johnny Ive, Realtime API
- (03)
Elon Musk vs OpenAI Trial, Google Cloud Surge, Meta’s Blocked Acquisition, Anthropic Winning
- (04)
Apple, OpenAI & Why Nobody Makes Money in AI | The SaaS Apocalypse
12:45 into “Apple, OpenAI & Why Nobody Makes Money in AI | The SaaS Apocalypse”
- (05)
OpenAI's $852 Billion Valuation Makes Zero Sense (ClawCon Tokyo, Anthropic's Leak, Goodbye Internet)
- (06)
Fire Sam Altman, The End of Software Engineers, and Why AI Is All Narrative
9:32 into “Fire Sam Altman, The End of Software Engineers, and Why AI Is All Narrative”
- (07)
Stripe Buys OpenRouter, OpenAI vs. Anthropic & the AI Backlash
33:22 into “Stripe Buys OpenRouter, OpenAI vs. Anthropic & the AI Backlash”
- (08)
Apple’s New Siri, SpaceX’s $300B IPO & the AI Company Killing Its Own Brand
5:51 into “Apple’s New Siri, SpaceX’s $300B IPO & the AI Company Killing Its Own Brand”
- (09)
Anthropic’s Bet on Coding Is Working (OpenAI Shopping Pivot, A16Z’s Top 50 List, $1B Tennis Channel)
- (10)
Silicon Valley Is Great at Tech and Terrible at Business
32:23 into “Silicon Valley Is Great at Tech and Terrible at Business”
- (11)
AI Frontier Lab Is Essentially Doomed | Open AI Offered U.S. Equity, Xbox Is In Trouble, Post-AGI
- (12)
The New Startup Stack: One Founder + Agents | Henrik Werdelin (Audos) and Ben Broca (Polsia)