The most binding constraint in North American data center development is no longer capital, chips, or land, it is power delivery, and specifically the multi-year interconnection queues standing between a signed lease and a first megawatt online.

In April 2026, PJM reopened its interconnection queue after a multi-year pause that began in 2022 due to a massive backlog. CBRE’s 2026 outlook names “power availability and certainty” as the top site-selection criterion, ahead of location and cost.

FERC’s December 2025 order clarified behind-the-meter colocation rules, which we read as an implicit acknowledgment that the traditional greenfield model cannot keep pace with AI-driven demand on its own.

One asset class has been sitting in plain view of nearly every DC developer, colocation, and infrastructure investor working in North America.

The US pulp and paper industry holds over 10GW of interconnected power capacity across ~220 operating mill sites, concentrated in the ISOs where new-build capacity is hardest to secure. For the vast majority, these are not brownfields in the loose sense of “land that once had industry on it.”

Many are active or recently-active industrial sites carrying the full stack a hyperscaler campus needs: high-voltage substations, onsite biomass and gas cogeneration, water rights, industrial zoning, and decades of community license to operate.

A handful of deals already point the way. Google’s 2009 acquisition of the Summa paper mill in Hamina, Finland now anchors €3.5B of cumulative investment and seven data centers. In July 2026, Aligned announced a $1.5B AI compute campus on the closed Pixelle Chillicothe integrated mill in Ohio.

Norske Skog Saugbrugs signed an LOI with Green Mountain in February 2026 for a colocation data center on the site of an operating SC magazine paper mill. EcoDataCenter is building 600MW on the former Stora Enso Kvarnsveden mill in Sweden. Amp Z has proposed a phased campus on the shuttered Southland Paper mill in Texas, targeting 2.1GW of capacity by 2029.

1) Why this pipeline matters more than it looks on paper

Start with geography. Based on AFRY’s analysis, the fleet skews heavily toward the Southeast (SERC/TVA), PJM, MISO, and MISO South, precisely the ISOs where hyperscaler demand is deepest and queues are longest.

Cost is the second driver. A meaningful subset of the fleet carries a low acquisition cost relative to greenfield alternatives. Mills under structural pressure, declining graphic paper demand, sub-scale assets, competitive margin compression, can often be sold or ground-leased for less than a DC developer would spend assembling equivalent land, interconnection, and water rights from scratch in the same ISO.

Then there is power itself. Integrated kraft mills with on-site generation typically runs 55 to 65 percent biomass (black liquor and wood residues), 15 to 20 percent natural gas, and only ten to 15 percent purchased grid electricity. Many sites are structurally net-neutral or net-export to the grid, with substations already sized for export. For a developer chasing behind-the-meter economics, a kraft mill’s 30-150MW cogen stack, maps almost exactly onto what a modern AI training campus needs.

And the ownership pool is tight: by AFRY’s count, ten operators control over 70 percent of the fleet by installed capacity. This concentrated universe is composed of a small number of professionalized industrial companies and their PE sponsors, many already running portfolio reviews and weighing non-core asset monetization.

In other words, the addressable universe is finite and knowable, and the decisions that matter will get made in a handful of C-suites over the next 24 to 36 months.

2) Where the bottleneck is worst, the opportunity is best

Not every pulp and paper site is equally attractive: a mill’s existing MW is worth roughly what it would cost to replicate that capacity greenfield in the same ISO. That is the scarcity-premium logic that has driven power markets for decades, now applied to industrial land.

In Virginia, Ohio, Pennsylvania, Indiana, and New York, where new build is hardest, a mill’s pre-secured interconnection can save five to eight years on delivery. Where new-build is easier, mills still carry real value, but the economics shift toward cogen retention and behind-the-meter structures rather than pure queue-jumping.

Across the Southeast belt, with its utility-negotiated interconnection paths and pro-development permitting, an integrated kraft mill’s black liquor and biomass CHP stack becomes the anchor of the commercial case rather than the interconnection date.

3) This is a relationships market, not a listings market

Here is the part that matters most: the overwhelming majority of the 10GW is not for sale. Mill owners will make monetization decisions on their own timeline, driven by their own strategic pressures.

So the winning play is not to canvass the fleet or wait for owners to come to market. It is to build advance conviction on which specific sites are worth pursuing, build the relationships and credibility ahead of time, and eventually approach P&P CEOs with a structured commercial proposition already priced against the mill’s operating alternatives.

That takes a specific analytical stack: pulp & paper cost curves and grade forecasts to identify which mills are structurally strained; ISO-level power market fundamentals to value the site’s power position; behind-the-meter economics to determine whether cogen retention or full acquisition is the right structure; and techno-commercial diligence to confirm the deal is bankable.

AFRY has built this framework, and the analytical tools supporting our conviction on this thesis. The Powered Mill is not a new opportunity: it is an overlooked one, hidden by an industry taxonomy that classifies these sites as manufacturing and misses their potential as powered industrial land. Whoever sees it first will move fastest.

AFRY advises large-scale pulp & paper producers globally, provides power market fundamentals via the BID3 platform to utilities, IPPs, hyperscaler PPA teams, and infrastructure funds, and holds a market intelligence partnership with DCD.