AI’s Infrastructure Bill Is Reaching Beyond Tech. One Midwest Factory’s Power Costs Rose 90%.

Belden Brick, a 141-year-old manufacturer in the Midwest, said its electricity costs rose 90% in 2025. Its monthly capacity charge, a fee tied to having enough power available when demand peaks, jumped from $1,600 to $12,000. The company makes bricks, not chips, but its rising power bill shows how the AI buildout is beginning to affect businesses far outside the technology sector.

The key observation is that the cost of a major investment boom does not always stay with the industry creating the demand.

Reuters reported on July 7 that Belden Brick’s electricity costs rose 90% in 2025. The company raised brick prices 4%, but profits still shrank.

The same report said capacity charges at Plaskolite, a plastics manufacturer with facilities in Pennsylvania and Ohio, rose from about $200,000 a year to $1.2 million.

Across the regional power grid serving much of the Mid-Atlantic and Midwest, capacity prices rose from $28.92 per megawatt-day in 2024 to $329.17 for the current delivery period.

The U.S. Energy Information Administration reported that average industrial electricity prices in April 2026 were 26% higher than a year earlier in Ohio and 14% higher in Pennsylvania. The national increase was 5.5%.

This is a cost-transfer environment.

AI data centers are adding large amounts of electricity demand to a power system that cannot expand overnight. New power plants, transmission lines, and other grid upgrades can take years to build.

When demand grows faster than supply, the higher cost may not stay with the companies building the data centers. It can spread to manufacturers and other large power users that depend on the same grid.

Bill Gates wrote a $100 million check.

Google signed a 15-year contract.

The Pentagon made it their top energy priority.

All for the same thing.

An energy source 140 times larger than global electricity demand. It runs around the clock. No fuel costs. No foreign supply chain. Zero emissions.

The problem was always access - it sits three miles underground, locked behind solid rock.

Last year a drilling crew solved that problem in 16 days. The government predicted 64.

Now Washington is handing this energy source an edge on August 18th that no competitor gets. Tax credits preserved while solar and wind lost theirs.

One company controls the technology. Sixty years of building. And the smartest money on Earth just showed up at their door.

See the company Gates, Google, and the Pentagon are backing >>

One key signal is the gap between industrial power prices in regions with heavy data-center growth and the national average. A wider gap can suggest that local power constraints are becoming a more serious business cost.

Another signal is how utilities and regulators divide the cost of new power plants and grid upgrades. The important question is not only how much electricity demand is growing. It is which customers end up paying for the system needed to support that growth.

A common misread is that higher factory power bills are purely an AI story. Data-center growth matters, but power-plant retirements, transmission limits, fuel costs, and slow additions to supply also affect electricity prices.

Another mistake is assuming that a major investment boom benefits every business around it. New spending can create jobs, construction, and demand while also raising the cost of scarce resources such as power, land, and labor.

The economic effects can move in both directions at once.

Focus on who absorbs the cost, not who creates the demand.

AI’s physical buildout is becoming easier to see beyond the technology sector. Demand for computing power is also demand for electricity, transmission, cooling, land, and new generation.

That makes the broader market frame less about whether AI demand is real. The more useful question is how the cost of supporting that demand moves through the economy.

For manufacturers, the difference may come down to pricing power. Some companies can pass part of a higher electricity bill to customers. Others may have to absorb more of it in their margins.

Large investment cycles often create gains at the center and costs at the edges. The rising power bills now showing up at Midwest factories are one sign that AI’s infrastructure needs are beginning to reshape the economics of businesses that have little to do with technology itself.

Talk soon,
The Playbook Daily