The technology industry is deploying artificial intelligence at a scale and speed that is outpacing the regulatory frameworks designed to govern it. SpaceX’s landmark IPO, banking regulators turning their attention to AI systems, and communities pushing back against data center siting are all symptoms of the same underlying shift: the AI economy is becoming a physical, infrastructure-heavy enterprise — and the world is only beginning to understand what that means.
The Infrastructure Imperative
Every major AI company is racing to build compute infrastructure that can support the next generation of models. Data centers that once housed servers quietly are now becoming industrial facilities the size of football fields, consuming as much electricity as small cities. The demand for power is so acute that some AI companies are negotiating directly with utility providers, bypassing the normal regulatory channels that govern grid expansion.
This infrastructure buildout is not confined to the United States. In Europe, where energy costs are higher and environmental regulations are stricter, the expansion of AI data centers has prompted political responses. The convergence of AI compute, satellite connectivity, and data center operations represents a new kind of industrial footprint that communities and regulators are only beginning to understand.
Banking Regulators Step In
U.S. banking regulators — including the Federal Reserve and the Office of the Comptroller of the Currency — are intensifying their scrutiny of how financial institutions deploy artificial intelligence. Unlike the fragmented, forward-looking approach that has characterized most AI regulation to date, these agencies are applying existing frameworks covering consumer protection, model risk management, and operational resilience to the AI systems banks are already running.
The focus areas are telling: lending decisions, fraud detection, compliance monitoring, and customer verification. These are precisely the use cases where AI has delivered the most immediate business value in financial services — and also the ones where algorithmic errors can have the most direct impact on people’s lives.
Financial services are among the most AI-dependent industries, making regulatory oversight here a preview of how AI governance may evolve across the broader economy. Regulators are signaling that the era of unsupervised AI deployment in high-stakes domains is ending.
The Physical AI Revolution
The technology industry is no longer content to run language models in data centers. It wants to build robots, launch satellites, redesign manufacturing, and reshape the energy grid. This is a qualitatively different ambition than the software-centric AI of five years ago, and it raises a correspondingly different set of questions about infrastructure, governance, and community impact.
Jeff Bezos is reportedly investing in what he calls an “artificial general engineer” capable of redesigning physical systems. Nvidia is developing specialized processors for AI workloads. A Barcelona robotics startup raised $85 million to build general-purpose factory robots. The common thread in every case is physical AI, not just software.
What Comes Next
The next twelve months will likely see the first significant legal and legislative challenges to AI data center siting in multiple jurisdictions. The normalization of biometric surveillance at major international events, the expansion of AI compute into physical infrastructure, and the concentration of AI power in a small number of vertically integrated platforms are all trends that will accelerate whether regulators are ready or not.
The SpaceX IPO is a milestone. What it represents is a turning point — the moment when the technology industry stopped being a software business and started becoming a physical one. The infrastructure it is building will shape communities, strain power grids, and draw regulatory attention for years to come. Understanding that trajectory is the job of this column, and it is the job we intend to keep doing.