Keep pulling the thread on Patrick Moorhead.
The AI industry is rapidly moving away from subsidized, flat-rate pricing models towards usage-based billing. This is driven by the immense token consumption of new agentic capabilities and the need for AI labs to achieve profitability, as exemplified by Anthropic's projected first profitable quarter.
AI models are beginning to solve long-standing, complex scientific problems that were previously intractable for humans. The instance of a general-purpose OpenAI model solving an 80-year-old mathematics problem is presented as a leading indicator of AI's potential to autonomously produce landmark results across various scientific fields.
The insatiable demand for computational power is driving a new phase of infrastructure development, with non-traditional players like SpaceX entering the market. Elon Musk is positioning SpaceX as a major AI compute provider, offering services at scale and floating futuristic concepts like orbital data centers.
Governments are actively trying to formulate policy responses to AI's rapid advancement, but the process is fraught with debate and uncertainty. This is seen in California's exploratory order on labor disruption and the White House's postponed federal order, reportedly derailed by concerns over stifling innovation in the race against China.
Major tech companies are moving beyond standalone AI chatbots to deeply integrate agentic capabilities into their core consumer products. Google's plan to embed persistent, information-gathering AI agents directly into Search exemplifies this trend, aiming to transform user interaction from one-time queries to ongoing, managed tasks.