Dylan Patel - The Infinite Demand for Tokens, Claude Mythos, and Supply Constraints
April 23, 2026
AI Summary
5 min readIn late 2024, Dylan Patel’s firm, SemiAnalysis, was spending tens of thousands of dollars a year on AI. By early 2025, that figure had exploded to a $7 million annual run rate—over 25% of the company’s $25 million salary expense. The catalyst was not a new business line or a hiring spree, but a single behavioral shift: employees, including non-technical ones, started using Claude Code to build production-grade tools. One person created a GPU-accelerated application that automates semiconductor reverse engineering—work that previously required an entire team at Intel. Another built a complete mapping of the U.S. power grid, scraping every power plant and transmission line, in a few weeks. A third, a former bank economist, single-handedly constructed a new benchmark of 2,000 economic tasks and a metric for measuring AI-driven deflation. Patel calls this phenomenon “Claude psychosis”—the moment when a user realizes that implementation difficulty has collapsed, and the only constraint is choosing which idea to execute.
The frontier model is the only model that matters
Continue reading the full summary in the app — free to try.
Read Full Summary →Free • No credit card required
Never miss an episode of Invest Like the Best with Patrick O'Shaughnessy
Get every new episode summarized in your inbox — free, ~5 minutes to read.
No spam. Unsubscribe anytime.
What you'll learn
- 1 (02:27) **Introducing Dylan Patel and the Token Economy Thesis** - Patrick introduces Dylan Patel, founder of Semi-Analysis, and sets up the core framework: the explosive demand for frontier-model tokens and the physical supply bottlenecks that constrain them.
- 2 (03:08) **Semi-Analysis’s Own AI Spend Explosion** - Dylan recounts how his firm's AI spending skyrocketed from tens of thousands to a $7 million run rate, driven by non-technical employees using Claude Code.
- 3 (05:15) **Concrete Use Cases: Chip Reverse Engineering and Economic Modeling** - Dylan gives two examples of how a few thousand dollars in tokens replaced entire teams and months of work.
- 4 (07:28) **Why a Business Owner Can’t Put the Brakes on AI Spend** - Dylan explains that in an information business, refusing to spend on the frontier model is existential because AI commoditizes everything.
- 5 (10:07) **Will Investment Firms Just Build It Themselves?** - Patrick asks if the ease of building will cause capital-rich investment firms to internalize all the data and insights.
- 6 (11:44) **What the Demand Explosion Teaches About Token Economics** - Dylan connects his firm’s experience to the macro picture, showing that Anthropic’s gross margins have soared to at least 72% because demand is so far above supply.
- 7 (14:55) **The Insatiable Hunger for the Frontier Model** - Patrick shares his own experience of instantly wanting the newest model (Opus 4.7) and being unable to use the previous one, and Dylan explains why this is rational.
+ Full timestamped outline available in the app
Guests on this episode
Show Notes
This is my second conversation with Dylan Patel. Dylan is the founder and CEO of SemiAnalysis, where he tracks the semiconductor supply chain and AI infrastructure buildout.
This conversation is about the supply and demand of tokens. On demand, Dylan describes something completely explosive. He explains why the frontier model is the only model anyone wants, and willingness to pay for it is nearly unbounded. His own firm has gone from tens of thousands of dollars in AI spend last year to seven million this year.
On supply, we walk through the bottlenecks across memory, logic, and fab equipment that will determine how fast any of this can scale.
We also cover Claude Mythos and what the leading labs need to do to fix their growing public perception problem.
For the full show notes, transcript, and links to mentioned content, check out the episode page here.
-----
Become a Colossus member to get our quarterly print magazine and private audio experience, including exclusive profiles and early access to select episodes. Subscribe at colossus.com/subscribe.
-----
Ramp’s mission is to help companies manage their spend in a way that reduces expenses and frees up time for teams to work on more valuable projects. Go to ramp.com/invest to sign up for free and get a $250 welcome bonus.
-----
Trusted by thousands of businesses, Vanta continuously monitors your security posture and streamlines audits so you can win enterprise deals and build customer trust without the traditional overhead. Visit vanta.com/invest.
-----
WorkOS is the infrastructure B2B and AI-native companies use to sell to enterprise. It covers everything enterprise security requires: SSO, SCIM, RBAC, Audit Logs, AI governance, and more. Trusted by 2,000+ fast-growing companies, including OpenAI, Anthropic, Cursor, and Vercel.
-----
Rogo is the AI platform for finance. They're building agents for Wall Street that are trained to understand how bankers and investors actually do work: from diligence and modeling,
More from this podcast
Invest Like the Best with Patrick O'Shaughnessy →