GeoMarketSignal

Yesterday, the Federal Reserve raised its policy rate by 25 basis points to 3.75%–4.00%. Yet the AI boom is still running. The technology is real, productivity gains may be emerging, and the investment surge is undeniable. The bubble question begins not with whether AI works, but with whether the expected return on the next AI project can keep exceeding the full cost of the capital required to build it.  Including myself, investors and market watchers have repeatedly compared today’s AI rally with the Internet boom of 1995–2000. The comparison is useful. But the way most people remember that period is too simple. The Internet boom was not a choice between real innovation and a financial bubble. It was both. Technology transformed the economy. Productivity accelerated, investment surged, and businesses poured capital into information technology and communications infrastructure. Financial excess developed alongside those genuine economic gains.   Research by Stephen Oliner and ...

  What NVIDIA’s August 26 Earnings Could Reveal About the Sustainability of the AI Investment Cycle NVIDIA will report its fiscal 2027 second-quarter results on August 26, 2026. For a company that has repeatedly broken revenue records, another earnings beat would hardly be surprising. NVIDIA has exceeded its own revenue guidance quarter after quarter, while demand for AI computing infrastructure has continued to expand at an extraordinary pace. But I believe this earnings report matters for a different reason. The question is no longer simply whether AI demand exists. The more important question is whether the earnings being generated by the AI ecosystem can continue growing fast enough to justify the enormous amount of capital required to build it, particularly when the price of that capital is rising. There are therefore two questions I will be asking when NVIDIA reports. Can NVIDIA outrun the rising cost of capital? And, more importantly: Can NVIDIA’s customers earn enough on AI...