Tokenomics: The Economy Of AI Part 1

This is the start of a multi-part series about memory chip stocks…but just like the SpaceX series began as a single article, and then evolved into a grand epic journey into the entirety of the human condition😉😂, well part 1 of this report became a delightful and very insightful deep dive into Tokenomics! Update On the AI Boom: Don’t Forget Sagan’s Law😉 Here is what I mean. RPO (remaining performance obligations) is the proxy for how much demand the hyperscalers have (contracted revenue) for AI datacenters. Source: FactSet Note the growth rate the company is reporting…363% in the last quarter, and even the slowest growth period back in 2025 was “just” 41%. Now compare that to the consensus of 13% CAGR growth through 2029. Estimates keep rising because analysts are Ok, so let me show you something really exciting. Let Me Explain A Simple But Powerful Model For Understanding the AI Boom Do you remember when DeepSeek came out in January of 2025 with R1, and we had “DeepSeek Monday”? Remember DeepSeek Monday? Jan 27th, 2025?😉 Remember how the AI boom was over because of DeepSeek's cheap model?😂 Jevon's Paradox is real...and Tech CEOs said so on DeepSeek Monday. They were right. DeepSeek claimed to have trained R1 on about $5 million BUT they didn’t include any costs other than the single training run. In other words, they distilled US models to create a training model, then set up an optimized training run for the final model, and THAT and only THAT run cost $5 million. The idea that a startup in China pulled a “Apple in a garage” and suddenly blew up the business models of Amazon, Google, and Microsoft, who were building AI datacenters to handle the data…(not training models themselves other than Gemini). Not to mention Jevon’s Paradox…which Satya Nadella (CEO of Microsoft) explained that very day. Jevon’s Paradox: The Heart Of How Technology Has Always Worked