A Huge Grid Charge Has Barely Moved Brick Prices

Belden Brick’s monthly grid-capacity charge jumped about 650%, from roughly $1,600 to $12,000, while the national producer price index for brick and structural clay tile rose only 0.82% from July 2025 to July 2026. The verdict is that AI data centers can raise a brick plant’s regional power costs, but the evidence does not show a nationwide brick-price shock—or even that Belden passed this charge into its selling prices.

The distinction is scale. Belden’s added capacity charge is about $10,400 a month, a serious plant expense. Spread over a plant’s annual output, however, that single charge can amount to less than one cent per brick. The available evidence does not provide Belden’s annual brick output, electricity’s share of its total costs, or a documented pass-through rate, so a precise effect on its prices cannot be calculated.

Enter a plant’s annual output and capacity charges; the result shows which side wins for those inputs.

Why The Construction-Cost Concern Sounds Plausible

The received view follows a familiar chain: AI data centers create very large, concentrated electrical loads; power supply and grid infrastructure fail to expand as quickly; utilities and regional markets recover the resulting costs; energy-intensive manufacturers pay more; and those manufacturers raise material prices.

That concern is not imaginary. A data center does not need to buy a single brick to affect a brick plant. Both can draw power from the same regional system. New loads can increase demand for generation, transmission, distribution and dependable capacity. Utility tariffs, customer contracts and regulatory decisions then determine who pays.

The Ohio evidence confirms part of this chain. Belden Brick in Sugarcreek reportedly faced a 90% increase in annual electricity costs, and its monthly PJM capacity charge rose from approximately $1,600 to $12,000. PJM’s capacity price rose from $28.92 per megawatt-day in 2024 to $329.17, a 1,038% increase, with rapidly growing data-center demand identified as a major contributor in the report on Belden Brick’s power costs.

As of the December referenced in that report, industrial electricity prices were up 7% nationwide, 26% in Ohio and 31% in Pennsylvania from a year earlier. A PJM representative said data centers can be completed faster than the generation needed to serve them, allowing demand to outrun supply.

The consensus is therefore right about the first two links: data-center growth can pressure a regional power system, and those costs can reach an existing brick plant. The Ohio case turns that mechanism from a theory into a documented operating expense.

What it does not establish is the final link. The reporting does not say Belden raised brick prices in response. It does not quantify electricity’s share of total production cost or isolate data centers as the sole cause of PJM’s capacity increase. Generation availability and other grid conditions also affect electricity markets.

A 650% Charge Increase Is Not A 650% Brick-Cost Increase

A percentage increase sounds dramatic without its starting weight in the finished product. The relevant relationship is: potential percentage pressure on brick prices equals the percentage change in electricity expense multiplied by electricity’s baseline share of total production cost and the share of the increase passed through.

Neither required share is available for Belden. The sources do not disclose the plant’s total cost structure or how much, if any, it would pass to customers. Kiln fuel, electricity, clay and other raw materials, labor, maintenance, overhead, packaging and freight all contribute to the delivered cost. The draft evidence supplies no percentage breakdown among them.

The capacity-charge arithmetic is narrower but useful. Moving from $1,600 to $12,000 adds $10,400 per month, or $124,800 over 12 months. Divide that annual increase by annual brick output to obtain the added capacity cost per brick. The calculator does exactly that; its default output is explicitly an illustrative mid-size volume, not Belden’s disclosed production.

This calculation should not be mistaken for Belden’s full electricity increase. The company reportedly experienced a 90% increase in annual electricity costs, but the source does not provide the annual dollar totals needed to convert that broader change into cents per brick. The calculator isolates only the reported monthly capacity charge.

A producer also has choices other than immediate, complete pass-through. It may absorb the expense through lower margins, change production schedules, reduce less-profitable output, delay expansion, improve efficiency, seek a different power arrangement or change only selected product and regional prices. Belden was reportedly exploring on-site electricity generation.

Those responses have costs and limitations, but they show why an input bill does not move mechanically into a price list. Contracts, competitive pressure, inventory and expectations about whether a charge will persist can delay or prevent pass-through.

The National Brick Index Barely Moved

The Bureau of Labor Statistics Producer Price Index for brick and structural clay tile provides the clearest national check. It stood at 269.785 in July 2025 and 272.008 in July 2026, a year-over-year rise of 0.82%, according to the BLS series presented through FRED.

The five most recent observations also show no sharp break:

The index was nearly unchanged from March through June and then edged down in July. That pattern is inconsistent with a broad national brick-price spike during the period when the regional capacity-charge story emerged.

This is a scoped finding. The index measures selling prices received by U.S. producers for the combined brick-and-structural-clay-tile category. Its base is December 1984, which equals 100; it is monthly and not seasonally adjusted. It is not a retail quote, delivered material price, contractor bid or installed masonry cost.

It also cannot isolate a state, manufacturer, color, size or product line. Freight, distributor margins and masonry labor fall outside it, and the series contains no variable identifying data-center demand or electricity costs. A local quote can rise while this national measure stays flat.

Cost and price timing can differ as well. A producer can carry higher utility costs under existing contracts, wait for a scheduled price-list revision or offset them operationally. The flat index does not prove that no plant felt pressure; it shows that a broad rise in producer selling prices had not appeared through July 2026.

The Lumber-And-Steel Analogy Breaks At The Cost Pathway

Coverage that treats data-center electricity as the next construction-material shock borrows the logic of tariff-driven lumber and steel stories. The analogy is limited. A tariff can strike the traded raw material or finished product directly. The PJM event strikes one component of a brick plant’s operating cost, after which the effect is diluted by the rest of the cost structure and by the producer’s response.

That does not make electricity trivial. Brick firing requires sustained heat, and a 90% annual electricity-cost increase can materially hurt a producer. It means only that the percentage change in one bill should not be presented as the percentage change in brick.

Historical masonry price movements also have other explanations. Gordian reported that common face-brick prices saw double-digit increases during the three months leading into its April 2024 update after a longer period of negligible change. Its concrete and masonry cost discussion cited raw-material or cement shortages, inconsistent supply, transportation costs, tariffs and general construction demand rather than AI data centers.

That comparison has its own limit because many of Gordian’s quantified updates concern concrete block. Clay brick and concrete masonry units use different principal materials and production processes; block prices are not a substitute for the brick PPI.

Direct Data-Center Demand For Brick Remains Unproven

The second possible pathway is physical demand: enough data centers specify enough masonry to strain nearby brick supply. Current evidence does not quantify brick or masonry units per data center, establish the prevalence of brick in their enclosure systems, or compare project demand with plant capacity and distributor inventory.

The total cost of a data center does not answer the material-quantity question. Turner & Townsend found that liquid-cooled facilities designed for AI workloads cost 7% to 10% more than comparable air-cooled facilities of similar IT capacity. Its indicative analysis assigned 9% of liquid-cooled facility cost to core-shell and architectural work, versus 14% for air-cooled facilities, while mechanical and electrical systems took much larger shares, according to its 2025 construction-cost analysis.

Those 9% and 14% figures are not masonry shares. Core-shell and architectural work contains many systems and materials, so the percentages cannot be converted into brick counts.

Cushman & Wakefield associates rising all-in development costs with power scarcity, land constraints, skilled-labor shortages, long equipment lead times and project complexity. Its 2026 development cost guide does not identify brick as a constrained input.

A local effect remains possible. Several masonry-heavy projects beginning near the same time could tighten a market with little spare production or inventory. That could lengthen lead times, limit colors or sizes, increase freight distances, change quotes or strain the masonry labor pool. Proving it would require project specifications, quantity takeoffs, schedules, regional output and inventory data. None are supplied by the available sources.

Regional Exposure Matters More Than National Data-Center Growth

An AI-related brick-cost risk is most credible where several conditions overlap: rapid data-center load growth, limited generation or transmission capacity, tariffs that expose existing industrial users, a brick plant subject to those charges, little spare manufacturing capacity and costly alternative supply routes.

Belden fits enough of that pattern to demonstrate exposure, not pass-through. It operates within a particular power market, utility arrangement and production system. A plant under different capacity rules, contracts or generation conditions could see a substantially different result.

Cost allocation is decisive. A large data center may bear more of the required generation and grid-upgrade expense under one tariff or special contract. Under another arrangement, some cost may reach other customer classes. A Harvard Law School discussion of data-center electricity costs describes that allocation as an empirical and regulatory question rather than a universal rule.

The first visible local effect may not be a factory price-list increase. It could be a delivered-price change, energy or freight surcharge, shorter quote-validity period, reduced product availability or longer lead time. Each still needs evidence before being attributed to AI; maintenance, transport disruptions and conventional construction demand can produce the same symptoms.

What Would Prove An AI Brick-Price Link

The electricity pathway needs a complete chain: documented data-center load growth; evidence that it contributed to a regional capacity or infrastructure cost; a tariff or bill showing that cost reached a brick manufacturer; the manufacturer’s total cost exposure; and a pricing notice or transaction record showing pass-through to buyers.

The Belden case reaches the manufacturer-cost stage. It does not reach the selling-price stage.

The direct-demand pathway requires different evidence: masonry takeoffs and enclosure specifications for data-center projects, their construction schedules, regional brick output, plant utilization, distributor inventories and available substitute supply. The number or dollar value of planned facilities alone cannot demonstrate a brick shortage.

For an actual masonry estimate, current comparable local quotes remain more useful than a speculative national AI surcharge. Separate the factory material price from freight, distributor charges and energy adjustments; confirm availability and quote duration; and review escalation language. A changed quote is evidence of a changed price, though its cause still has to be established.

The bounded answer remains: AI data centers have demonstrably contributed to a severe regional grid-cost problem that reached at least one brick maker. Through July 2026, the national brick and structural clay tile index moved just 0.82% year over year, and no sourced evidence shows Belden passing its higher capacity charge into brick prices.