Episode summary
The Supply and Demand of AI Tokens | Dylan Patel Interview
Invest Like the Best
· · 46 min · Hosts: Patrick O'Shaughnessy · Guests: Dylan Patel
Summary
Dylan Patel argues that AI has moved from a world where execution was scarce and ideas were cheap to one where implementation is cheap and the best ideas are what justify enormous token spend. At SemiAnalysis, that shift is already visible: the firm’s annualized spend on Claude and related tools has risen from tens of thousands of dollars to about $7 million a year, or more than 25% of salary expense, because nontechnical staff and engineers alike are using AI to build software, reverse-engineering tools, and large data products.
He says the real story is not just cheaper software creation but explosive demand for frontier models and the willingness of users to pay for them. Patel claims Anthropic’s revenue has surged because customers want the newest and best models, even at higher prices and tighter rate limits, and that many businesses will be priced out unless they can turn tokens into economic value. He expects this dynamic to intensify as model quality improves and access becomes more concentrated among large, well-connected customers.
On the supply side, Patel argues the bottleneck has shifted across the whole stack: GPUs, memory, CPUs, optics, copper foil, and especially TSMC capacity. He says lead times are long, supply is sold out, and capex keeps rising, while useful GPU life is stretching. He also thinks robotics will become a meaningful new demand curve and that the social reaction to AI will eventually turn into large-scale protests.
Key takeaways
- Patel says SemiAnalysis now spends about $7 million a year on cloud code, up from tens of thousands, because AI can compress whole workflows into a few thousand dollars of tokens.
- He argues the key business test is whether token spend creates more revenue than it costs; otherwise, firms will get commoditized and priced out.
- Patel says Anthropic’s demand is so strong that users care more about rate-limit access and enterprise contracts than list pricing.
- On supply, he says GPU useful life is stretching and bottlenecks now include memory, CPUs, TSMC capacity, and even copper foil and PCB inputs.
- Patel expects robotics to add another major demand curve, but thinks the bigger near-term risk is backlash and large-scale protests against AI.
Notable quotes
“What used to matter a lot was execution was very very [ __ ] difficult and ideas were cheap”
“Now ideas are cheap and plentiful but execution is very easy”
“I think when I last talked to you it was 5 million spend rate. It's actually 7 million spend right now”
“if I don't move fast enough I will also lose my edge so the question is yes AI commoditizes things just like it commoditizes software”
“what really matters is having an anthropic rep and having an enterprise contract with them and getting the rate limit increases that you need”
How the episode opens
What used to matter a lot was execution was very very difficult and ideas were cheap. Now ideas are cheap and plentiful but execution is very easy. So really only the good ideas are the ones that can justify the spend on super cheap implementation.
You told me this incredible story about how your own team's use of tokens has changed dramatically this year. Yeah. Retell that story and what it is teaching you about what's going on in the world.
Last year we thought we were heavy users of AI. Everyone's using chat GPT. Everyone's using cloud.
Everyone's got you know I'm providing whatever subscriptions anyone wants on the order of spend of like tens of thousands of dollars for our firm. This year the spend has just skyrocketed and and it really started in late December with Opus that included Doug who's president uh Doug Olaflin. He's very much like leading the charge in the sense of like non-technical people using uh AI for coding.
Um and so he's basically pled the whole firm slowly over time. I think he's been the the leader in doing that. Obviously the engineers were using it anyways but spend in January just started to inflect and rocket and rocket and rocket and rocket.
Um, we signed, you know, an enterprise contract with Anthropic and it's gone to the point where now, um, I think when I last talked to you it was 5 million spend rate. It's actually 7 million spend right now.
That was last week, by the way. — A lot of that is just the usage, right? What's what's really, you know, people people who are have never coded before are using cloud code and spending thousands of dollars sometimes a day, but across a firm, we're spending $7 million a year now on cloud code at the current rate.
From the episode’s captions: roughly the first two minutes, lightly cleaned. Keep watching on YouTube →
Topics
- ai token supply and demand
- ai capex and compute bottlenecks
- semiconductor supply chain
- model hoarding
- public backlash against ai
Companies and tickers mentioned
- TSM
- SemiAnalysis
- Anthropic
- OpenAI
- TSMC
Watch the episode
Watch “The Supply and Demand of AI Tokens | Dylan Patel Interview” on YouTube
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from Invest Like the Best
Invest Like the Best, this episode and its artwork belong to Colossus | Investing & Business Podcasts. Octagon is not affiliated with the show; this page is an independent summary.
