We scan new podcasts and send you the top 5 insights daily.
Critiques of "circular financing" in AI (tech giants funding startups who buy their products) miss the point. This is simply efficient capital deployment to meet real demand. The key test is whether the compute capacity is fully utilized by end-users with positive ROI applications. With no "dark GPUs" in the market, this concern is currently unfounded.
The dot-com crash was fueled by massive overinvestment in infrastructure (dark fiber) with no corresponding demand. Today's AI boom is different: every dollar spent on GPUs has immediate, pent-up customer demand, making the investment cycle fundamentally more sound.
Unlike the dot-com bubble's speculative fiber build-out which resulted in unused "dark fiber," today's AI infrastructure boom sees immediate utilization of every GPU. This signals that the massive investment is driven by tangible, present demand for AI computation, not future speculation.
The current AI infrastructure buildout, while massive, is fundamentally different from the dot-com bubble. It's financed by cash flows from highly profitable companies, not speculative debt. Crucially, demand is real and immediate; unlike the 'dark fiber' of the 90s, there are 'no dark GPUs' today.
Martin Shkreli reframes the critique of circular AI investments (e.g., Nvidia invests in OpenAI, which pays Oracle, which buys Nvidia chips). He argues this isn't a flaw but simply an "economy." Its legitimacy is proven not by internal transactions, but by the strong and growing demand from outside users and companies.
The AI ecosystem appears to have circular cash flows. For example, Microsoft invests billions in OpenAI, which then uses that money to pay Microsoft for compute services. This creates revenue for Microsoft while funding OpenAI, but it raises investor concerns about how much organic, external demand truly exists for these costly services.
The current AI infrastructure expansion differs critically from the dot-com bubble's fiber buildout. There are no 'dark GPUs'; every unit of computing power, even older generations, is immediately utilized, suggesting demand is keeping pace with supply.
The comparison of the AI hardware buildout to the dot-com "dark fiber" bubble is flawed because there are no "dark GPUs"—all compute is being used. As hardware efficiency improves and token costs fall (Jevons paradox), it will unlock countless new AI applications, ensuring that demand continues to absorb all available supply.
Unlike the dot-com era where capital built unused "dark fiber," today's AI funding boom is different. Every dollar spent on GPUs is immediately consumed due to insatiable demand. This prevents a supply overhang, making the "circular funding" model more sustainable for now.
The current AI infrastructure build-out avoids the dot-com bubble's waste. In 2000, 97% of telecom fiber was unused ('dark'). Today, all GPUs are actively utilized, and the largest investors (big tech) are seeing positive returns on their capital, indicating real demand and value creation.
Unlike sham transactions that invent revenue, investments like Nvidia's into its GPU customers are economically sound. The deciding factor is the massive, verifiable downstream demand for the AI tokens these GPUs produce. This makes the deals a form of strategic credit extension, not fraudulent accounting.