On June 18, the Federal Energy Regulatory Commission did something it almost never does. It issued show-cause orders to all six US regional grid operators, telling each one: justify your current rules for connecting large power loads, or change them. The target was obvious. AI data centers.

The five commissioners voted unanimously. PJM Interconnection, MISO, Southwest Power Pool, CAISO, ISO New England, and NYISO all received tailored orders. Grid operators have 60 days to respond. They also have 30 days to report how much spare generating capacity they have left, if any.

The stated reason: AI-driven power demand is growing faster than the grid’s interconnection process can handle. Companies that want to build data centers are stuck in queues that were designed for a world where large new loads appeared once or twice a year. That world ended. FERC decided the rules need to catch up.

This is not a policy signal. It is a structural one.

When a federal regulator forces six regional grid operators to restructure simultaneously, it is not sending a message. It is changing the ground rules. The companies that already have sites permitted, land acquired, and power purchase agreements in progress just got a regulatory tailwind. The companies still debating whether AI is real enough to invest in did not.

That distinction matters more than any model release this month.

The directive traces back to Energy Secretary Chris Wright, who asked FERC in October 2025 to consider reforms for connecting large loads faster. Eight months later, the commission moved from “consider” to “justify or change.” That acceleration tells you something about how seriously the federal government takes the infrastructure bottleneck. They are not studying the problem anymore. They are forcing a solution.

The infrastructure gap is now a policy gap

A month ago, NextEra Energy announced a $67 billion acquisition of Dominion Energy. The largest utility merger in American history, driven by AI power demand projections. Now the federal government is restructuring interconnection rules to match.

These are not independent events. They are the same conclusion reached by different institutions: AI infrastructure is not a technology trend. It is a load-bearing economic shift. The power grid, the regulatory framework, and the capital markets are all reorganizing around it at the same time.

For the companies already in motion, this is compounding advantage. They secured land when it was available. They locked power purchase agreements before demand spiked. And now the federal government is clearing the regulatory path they were already on.

For everyone else, each of those advantages is harder to replicate with every passing quarter.

What this means if you run a business

You do not need to build a data center to feel this. The downstream effects hit every company that depends on AI infrastructure, which increasingly means every company.

Start with cloud. Faster grid connections mean faster capacity expansion, which means the hyperscalers can bring new regions online sooner. If your AI workloads run on AWS, Azure, or Google Cloud, you benefit indirectly from FERC’s action. But only if you are already running workloads. If you are still in the planning phase, you are relying on capacity that your competitors are already consuming.

Then there is energy pricing. FERC explicitly required grid operators to protect existing customers from cost shifts. But “protect” in regulatory language means “manage the transition,” not “prevent any change.” When 300-megawatt data centers connect to grids designed for 50-megawatt peak loads, pricing structures change. Businesses that depend on stable energy costs should be watching this closely.

And talent. Every new data center that comes online faster needs engineers, operations staff, and security teams. The same people your AI projects need. The companies building infrastructure are pulling from the same labor pool, and they are offering infrastructure-scale compensation to do it.

The real question FERC answered

The commission’s five categories of inquiry cover technical details: interconnection study procedures, cost allocation, affected system coordination, surplus capacity, and transparency. But the real question FERC answered is simpler than any of those.

Is AI infrastructure permanent enough to restructure the grid around?

The answer, delivered unanimously by five federal commissioners, is yes.

That answer matters outside of power policy. When the institution responsible for the nation’s electrical reliability concludes that AI infrastructure is a permanent load on the grid, it removes one of the last credible arguments for waiting. The “let’s see if this is real” position just lost its strongest remaining refuge.

Six months from now, the grid operators will have responded and data centers stuck in multi-year queues will be breaking ground. The distance between organizations that moved and organizations that watched will be wider than it is today. As of June 18, the federal government is paving one side.