Power Grids Are The New Silicon

Power Grids Are The New Silicon

Discourse around AI compute is often framed in terms of vertical integration and the risk of foreclosure. But market realities may not align with these concerns. The main problem is not access, but the slow pace of capacity expansion due to power and permitting constraints. This conversation reframes the AI race as a race of infrastructure scarcity and execution speed. Read the full article and graphic analysis: https://blog.econworks.com/p/competition-in-ai-compute-is-about?r=562wri [https://blog.econworks.com/p/competition-in-ai-compute-is-about?r=562wri] Explore more visual economics content: https://econworks.com YouTube: https://www.youtube.com/@EconWorks-d3e Substack: https://blog.econworks.com [https://blog.econworks.com] This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit blog.econworks.com/subscribe [https://blog.econworks.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2]

The Deal That Almost Gave Meta Control Over AI Agents

This installment of Control in AI takes a look at the Meta–Manus transaction—a deal that was signed, entered into a phase of early technical exploration and alignment, but never reached the stage of full operational control. We discuss why the deal creates a new and important category in AI markets: partially realized deals. Such arrangements can impact technical direction, system design, and the nature of competition well before any formal change of control. We also link the case to broader questions around OpenAI and discuss why traditional antitrust frameworks struggle with the layered nature of AI (attention vs. execution layers) and why the central question is moving from “Is this deal a merger?” “When does the control actually begin?” This analysis is clear and practical for anyone interested in AI, antitrust, and platform power. Read the full article and graphic analysis: https://blog.econworks.com/p/control-in-ai-episode-3-meta-manus?r=562wri [https://blog.econworks.com/p/control-in-ai-episode-3-meta-manus?r=562wri] Explore more visual economics content: https://econworks.com YouTube: https://www.youtube.com/@EconWorks-d3e Substack: https://blog.econworks.com [https://blog.econworks.com] This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit blog.econworks.com/subscribe [https://blog.econworks.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2]

The Subsidy Paradox — Why Cursor Sold to SpaceX

We break down the economic structure of the AI developer market in this episode. We discuss how dependence on upstream foundation models compels downstream innovators to flow both capital and highly refined telemetry back to their vertically integrated competitors. We objectively evaluate the proposed solution, the acquisition of Cursor by SpaceX/xAI. Is the deal a classic case of vertical foreclosure that will harm competition or a necessary defensive merger that ends a data monopoly? We walk through the specific efficiencies of the deal, including cost reduction, performance optimization, and the removal of the innovation barriers. Read the full article and graphic analysis: https://blog.econworks.com/p/the-subsidy-paradox-why-cursor-sold?r=562wri [https://blog.econworks.com/p/the-subsidy-paradox-why-cursor-sold?r=562wri] Explore more visual economics content: https://econworks.com YouTube: https://www.youtube.com/@EconWorks-d3e Substack: https://blog.econworks.com [https://blog.econworks.com] This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit blog.econworks.com/subscribe [https://blog.econworks.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2]

Control in AI—Episode 2: Governance and Control in AI

AI Antitrust Series. Episode 2: In our second episode, we deconstruct the AI stack (Model → System → Interface) and show how control at the model layer determines access, capabilities, integration, and ultimately who wins the AI economy. We explore why economies of scale, learning effects, and scope make the model layer the main bottleneck. We compare it to the Apple ecosystem and discuss the two competing visions: AI as open infrastructure vs. AI as a tightly controlled platform. By the end of this episode, you will understand why governance is the new question at the heart of AI—not just who gets to build the best model but who gets to decide the rules of the whole system. It’s perfect for anyone who wants to understand the deeper economics and power dynamics behind the AI revolution. Read the full article and graphic analysis: Explore more visual economics content: https://econworks.com YouTube: https://www.youtube.com/@EconWorks-d3e Substack: This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit blog.econworks.com/subscribe [https://blog.econworks.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2]

Are Platforms Breaking Antitrust?

What happens when the “market” isn’t a product—but a platform? In this episode, we unpack why traditional antitrust tools struggle with platforms and introduce a new way to think about them. From credit cards to Google to sports leagues, not all platforms work the same way—and that distinction matters more than you think. Read the full article and graphic analysis: https://blog.econworks.com/p/when-the-platform-is-the-market?r=562wri [https://blog.econworks.com/p/when-the-platform-is-the-market?r=562wri] Explore more visual economics content: https://econworks.com YouTube: https://www.youtube.com/@EconWorks-d3e Substack: This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit blog.econworks.com/subscribe [https://blog.econworks.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2]