The “Project River” project in Charleston, Bradley County, Tennessee proposes a 500 megawatt (MW) biomass power plant that would also produce biochar, as well as a 1,000 MW gas plant, to power what the Hiwassee River Alliance expects will be an AI data center. This is the second data center proposed for this community*, located in the Hiwassee River watershed, which stands out nationally for the amount of pollution it already endures.
* “Project Midspan” wants to build a 300 acre datacenter about two miles from the proposed Project River site, and is suing Bradley County over its ordinance putting limits on datacenter location.
For context, 1,500 MW can power more than 800,000 single-family homes in the region simultaneously (based on Tennessee’s average home use of 1,154 kWh a month, EIA); Bradley County has 42,596 households (U.S. Census Bureau). Powerplants producing that much electricity would emit huge amounts of air pollution and CO2 and consume massive amounts of resources. The developer’s website downplays these impacts.
The Southern Environmental Law Center (SELC) has raised concerns about the environmental impacts of the project and of the process by which it is being reviewed. In its letter to city officials, SELC cites the risk of “extraordinarily harmful impacts” on residents and raises several feasibility issues should Project River be pursued as planned.
We share SELC’s concerns and support their analysis. Below, we expand on it with additional data and analyses about the “woodshed” and the pre-existing environmental conditions in the region. We also provide interactive visuals for the interested public and media to explore based on publicly available data.
The biomass plant, by the numbers:
Below, we explore some of the developers’ claims.
The developers have said the 500 MW biomass plant would use 5 million tons of Southern pine a year, but say 6 million tons on their website. But both figures underestimate actual demand. To provide 500 MW of electricity, assuming normal downtime and a typical efficiency of 24%, the plant would need about 6.7 million green tons of wood a year at the industry-standard 50% moisture, 33% higher than the lower estimate. The plant would demand even more wood if it is also producing biochar.
The developers told reporters their wood is “approximately 40% moisture.” But freshly cut pine is about half water by weight; 50% is the industry standard. Their own website says the same thing, describing the wood’s moisture as “about half its weight.” Getting this wrong isn’t a small slip. Wetter wood delivers less energy and material per ton, so every ton of 50%-moisture wood makes less electricity and less biochar than they’re counting on. It suggests they don’t understand the basic realities of harvesting and burning wood.
The developers say the plant will burn pine sourced within a 200-mile radius (and in some places they’ve said a 250-mile radius). But consistently hauling low-quality wood over 100 miles to a mill is rare. In one survey, loggers reported an average haul distance of 50 miles, which matches industry data. Hauling costs quickly outrun the value of the wood, which is why mills draw most of their wood from much closer.
Figure 1. Diesel particulate matter in the air. The highest tract near the site has 0.49 µg/m³. Colors follow EnviroAtlas’s own ranges, where yellow is lowest and green is the middle range.
While the economics of hauling wood over long distances seem implausible, close to the site, there isn’t much pine, and the pine that exists is already being cut. Forests within 100 miles are mostly oak–hickory hardwood; pine-type forest is only 13% of the 11.2 million acres of timberland in that radius. Loggers in that region already cut about 2.7 million green tons of pine a year for existing mills (see map below). Project River would demand 6.7 million tons on top of that:
Figure 2. Demand for pine from Project River as compared to current demand for pine in the woodshed at 50–200 miles from the proposed site.
Looking beyond the 100-mile radius does find more pine, but it finds it far away and already spoken for. Within 200 miles, Project River would add 20% to pine logging, but most of that pine is in the Georgia and Alabama pine belt. That is exactly where existing mills are thickest: 245 mills within 200 miles already buy pine, including 6 other wood-burning energy plants and pellet mills. The developers don’t acknowledge any of them.
In addition to the competition within the pine market, not all wood is available in the area to begin with. The Cherokee National Forest falls within the buying radius, but logging in National Forests requires additional contracts and longer processes. Private landowners may also be unwilling to sell timber, instead valuing their properties for natural beauty, hunting, and something to pass onto their kids.
Darker areas cut more pine per square mile each year. Orange dots are mills that already buy pine, and white dots are nearby hardwood mills. Rings mark 50, 75, 100, 150 and 200 miles from the site. Use the switch at top right to see individual counties instead of distance zones. Hover for values and sample sizes.
Figure 3. Pine harvest by distance zone, with existing mills.
The developers have not publicly disclosed engineering plans for the biomass plant. However, a poster presented at a meeting on Sept. 1, 2026, and an earlier version of the website (archived on the Wayback Machine on Sept. 8, 2026) indicate that rather than using direct combustion to burn wood and generate electricity, the 500 MW biomass plant will use “pyrolysis,” a technology that heats wood in the absence of oxygen. Products include “syngas” which is burned in a boiler to generate steam that turns a turbine that generates electricity, and “biochar” (essentially charcoal). They claim production of biochar makes the project “carbon negative” (more on this claim, below).
Pyrolysis exists, but it’s still being optimized as a technology for generating energy and products. There are pyrolysis plants being developed around the world that produce a few tens of thousands of tons of biochar, but no plants generating and burning syngas at anywhere near the scale of the 500 MW plant proposed for Bradley County.
It’s not clear why they want to use pyrolysis and produce hundreds of thousands of tons of charcoal. The entire US market of biochar in 2025 was about 151,329 metric tonnes (167,000 US tons), a fraction of the amount Project River plant could produce assuming the technology were even feasible at this scale. If the goal of the biomass plant is to generate electricity, then the most straightforward way to do this is with conventional combustion, a technology that has changed little since the early 1900s. The wood use numbers the developer has posted are consistent with assuming conventional combustion.
The developers claim the project is “cleaner by design”, but cramming two massive power plants plus a datacenter together in a small area means the region would be a hotbed of pollution. The Southern Environmental Law Center estimates the two powerplants alone would annually emit:
Burning wood is extremely polluting, and even with modern emission control equipment, biomass power plants tend to be disproportionately polluting for their size compared to same-sized fossil-fired plants with similar emissions controls. As there do not exist any syngas-burning power plants of comparable size to the 500 MW plant proposed at Project River, and there is no public information on what emission control technologies the plant would use, it is not possible to estimate how much pollution the plant would emit if it used this technology.
However, it is possible to examine air permits from other plants to estimate how much pollution the plant would emit if it used direct combustion. SELC developed their estimates by upscaling emissions from other air permits. This analysis independently estimates what emissions could be solely for the biomass plant, also based on previous permits. The following estimates do not include the emissions of volatile organic compounds from drying the wood, which SELC estimates in their letter.
Criteria pollutants and HCl: The “Trees, Trash and Toxics” report reviewed air permits from more than 80 biomass plants. Taking the highest and lowest permitted rates for the four largest plants in the database (all of which had boiler capacities greater than 1,000 MMBtu/hr and employed modern emission controls), and applying these rates to a 7,108 MMBtu/hr boiler (or multiple boilers summing to that figure), we estimate the Project River biomass plant alone could emit:
These estimates assume steady-state operation. As pollution emission rates increase during startup and shutdown (see page 27 of the “Trees, Trash, and Toxics” report), the more a plant cycles, the more pollution it emits.
Hazardous air pollutants (HAPs) include hydrochloric acid (HCl), toxic volatile organic compounds like benzene and formaldehyde, polycyclic aromatic hydrocarbons (PAHs), and heavy metals, including lead, cadmium, and mercury. Even clean, “natural” wood can be a surprisingly large source of heavy metals, because trees can accumulate metals in soils and the air that have been released by coal-burning and other industrial activities. Assuming typical HAPs emissions factors from the industrial lobbying group NCASI, as well as the EPA AP-42 database, a 500 MW wood-burning boiler could emit between 237 tons and 1,185 tons of hazardous air pollutants a year.
The project website downplays risks in a way that’s completely unjustified considering the developers have not publicly released any engineering plans. Regarding noise, the website claims the project will be a “quiet neighbor.” In reality, the biomass plant alone has the potential to be extremely loud, never mind the combined impact of a 500 MW biomass plant, a screaming data center, and a 1,000 MW plant. Intolerable noise was a big problem for the 102.5 MW (gross 116 MW) biomass plant built in Gainesville, FL, where residents described the sound as being like the roar of a jet engine (see p. 17 of PFPI’s report on how Federal Stimulus money was wasted on bioenergy). The plant’s wood piles and operations also created strong odors that could be smelled from up to two miles away. That plant also had a fire at its wood-processing unit. Fires are common at biomass plants, sometimes occurring due to fermentation and spontaneous combustion in woodchip piles (see p. 11 of the report and search “biomass plant fires” on the internet).
EPA data show Project River would be in one of the most polluted areas in the entire United States.
Figure 1 above shows ambient air toxics from diesel. As Figure 4 shows, cancer risk from air toxics overall is elevated in the region, as well as respiratory risk (Figure 5) and non-cancer neurological risk (Figure 6). (All the ensuing figures are from EPA’s EnviroAtlas interactive map).
How to read these maps: colors show how each census tract or watershed compares with every other one in the country; this is our ranking of EPA’s values, not an EPA rating. The dark outline marks the 4 census tracts within 2 miles of the site, or the site’s watershed. Hover over any area for its value. Every map opens at the same view; zoom out to see the rest of the region within 50 miles. Air toxics are from EPA’s 2014 National Air Toxics Assessment, metals discharges from 2011 permit reports, and mercury listings from 2015–16.
Figure 4. Cancer risk from all air toxics: 41 in a million near the site, higher than 87% of US tracts.
Figure 5. Respiratory hazard index from all air toxics. Both Bradley County tracts beside the site are above 1 (1.04 at the highest), the level at which EPA considers harm possible. Only 99 of 73,450 US tracts (0.13%) reach that level, and these two are the highest within 50 miles.
Figure 6. Neurological hazard index from all air toxics: 0.155 in the tract just west of the site, in the top 0.8% of US tracts and #3 of 272 within 50 miles (still below EPA’s level of concern, 1).
The watershed (designated by its hydrological unit code, “HUC”) is already the dumping ground for very large amounts of toxic metals (Figure 7). As zooming out the map shows, this watershed is literally unique in the region for the amount of permitted metals discharged, and rare even in the US for the total amount of permitted discharges. Mercury is a main source of impairment of the rivers around the project site (Figure 8).
Figure 7. Permitted metals discharges: 514,972 toxic-weighted pound-equivalents a year in the site’s watershed, the most of any watershed within 50 miles and in the top 0.1% of the 19,946 US watersheds with reported discharges (2011 data).
Figure 8. Streams impaired by mercury: 15.5 km in the site’s watershed, higher than 98% of US watersheds.
The website claims that “the biomass loop is carbon-negative” (implying it pulls CO2 out of the atmosphere) because “the biomass side holds back more carbon than it releases,” meaning some wood carbon would be retained in biochar. This claim is simply wrong. In reality, logging and burning wood emits massive amounts of CO2, and retaining some carbon in solid form as charcoal does not change that fact. Yes, forest growth pulls CO2 out of the atmosphere, but turning wood into charcoal does not.
The Project River biomass plant as described would emit more than 6 million tons of CO2 per year. If the plant did make biochar, some of the carbon in the wood would be retained, not emitted. But because the carbon retained in biochar isn’t converted to electricity, generating 500 MW of electricity would thus require burning more trees per megawatt-hour, not fewer. The developers haven’t said how much biochar the plant would make, or where it would go.
Emissions: At 50% moisture content, burning or pyrolyzing 6.7 million green tons of wood releases about 6.1 million tons of CO2 a year from the smokestack (subtracting out the water, bone dry wood is about 50% carbon; each ton of carbon becomes 3.67 tons of CO2). Trucking adds about 154,000 tons. (SELC estimated the biomass plant would emit 4.97 million tons of CO2 emissions based on EPA’s CO2 emission factor and the developer’s estimate of 5.5 million tons of wood.)
The developers say the plant will “make its own water from moisture recovered as the wood dries,” about 110 million gallons a year (per a developer poster, as reported in SELC’s letter). It’s not clear what technology they are planning to use. Closed loop cooling does massively reduce water use compared to open-loop cooling, but requires infrastructure and more electricity to run the equipment.
However, the developers are talking about getting water directly from wood. How feasible is this?
Each row assumes all of the plant’s wood comes from one distance band; the last row spreads it across the whole 200-mile area in proportion to where the pine is.