U.S. data centre power demand is projected to surge 755% by 2029, creating a 57-gigawatt supply gap that reframes the entire AI infrastructure investment thesis. Plus: AMD acquires World Labs for $8.2B, Nvidia authorises a $150B buyback at its lowest valuation in a decade, and Samsung foundry sets its sights on TSMC.
Audio is available on Spreaker — see link below.
The single most important shift in AI infrastructure right now isn't about chips. It's about electricity.
The response from hyperscalers is already taking shape, and it's not waiting for the grid. Behind-the-meter gas generation, meaning power produced on-site rather than drawn from the utility grid, is projected to hit twenty-nine point six gigawatts by twenty thirty.
While the power story was developing, AMD made its largest strategic move in years. The company acquired World Labs in an all-stock deal valued at eight point two billion dollars.
Nvidia's management sent its own signal this week. The company announced a one hundred fifty billion dollar share buyback authorization, bringing total cumulative authorization to two hundred thirty-five billion dollars.
Samsung's foundry division is targeting sixty-six percent of revenue from high-performance computing by twenty twenty-nine, up from twenty-eight percent today. The company's SF1.4 process node, at one point four nanometers, is slated for mass production that same year.
On the model and application side, OpenAI's DevDay this week launched Dots, its personal AI agent product, running on the GPT-6.1 Sol model with a safety framework called OpenShell. Dots is live for Pro and Business users in eligible markets.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.