Can AI Keep Growing Exponentially? Let’s Ask SemiAnalysis — With Dylan Patel and Jordan Nanos

The AI infrastructure build-out is currently consuming roughly 5-6% of U.S. GDP, driven by massive capital expenditures that far exceed historical physical infrastructure cycles. While market analysts anticipate decelerating growth, the demand for AI-driven knowledge work remains robust, with major labs like Anthropic and OpenAI moving toward profitability as their compute-to-revenue ratios improve. Nvidia plays a critical role in this ecosystem by providing significant financial backstops for neocloud providers, though these shared clusters face severe cybersecurity risks that threaten the integrity of sensitive research. As the industry matures, frontier AI labs are increasingly keeping their most advanced models internal to protect competitive advantages and capture higher value, signaling a shift from open API-based models to proprietary, high-margin enterprise applications that automate complex tasks across industries like drug discovery and chip design.