A large-scale data center campus known as Project Double Reed is planned for the STAMP industrial site in Genesee County. According to the applicant’s February 18, 2026 SEQR filing, the project would occupy roughly 90 acres, divided between a 60-acre North Campus and a 30-acre South Campus. It is designed as an extremely power-intensive facility: the project’s utility filing says it is expected to require about 500 megawatts of electric load. In comparison, STAMP’s electric plan contemplates a substation system expandable to 600 megawatts.
The project summary submitted to GCEDC also states that the developer is seeking a package of public assistance, including major sales-tax and mortgage-recording-tax exemptions, and a long-term PILOT arrangement, while projecting substantial electricity-related tax payments and other revenues over time. In short, this is not a conventional industrial project with large permanent employment; it is a very large, utility-intensive data center proposal whose economic case depends heavily on assumptions about long-run operating benefits and public costs. Are the sponsor’s claims well supported?
Media coverage has focused on several recurring concerns about the STAMP data center project: the project’s enormous scale — 2.2 million square feet and about 500 megawatts of power demand — compared with its relatively small permanent employment base of roughly 120 to 125 jobs; the size of the requested public subsidy package, described by Investigative Post as amounting to millions of dollars per job and by another Investigative Post report as among the largest subsidy requests in state history; the risk that the facility could raise electric rates or strain the power system; and broader concerns about water, wastewater, environmental impacts, and the project’s location near sensitive natural and cultural resources, as reflected in coverage by WKBW, The Business Times, and Orleans Hub.
Another issue raised prominently in the press is governance: Investigative Post reported criticism that the Genesee County IDA/GCEDC may face a conflict of interest because it stands to receive substantial fee income if the project is approved. Overall, the media has portrayed the proposal less as a straightforward economic development success and more as a contested trade-off involving subsidies, infrastructure burdens, environmental impacts, and limited long-term job creation.
Government Assistance Requested
The sponsor is seeking roughly $1.46 billion in total state and local tax subsidies. According to the March 12, 2026, Genesee County IDA/MRB cost-benefit summary, the request includes about $715.9 million in local sales-tax exemptions, $715.9 million in state sales-tax exemptions, $15.6 million in local mortgage-recording-tax exemptions, and $15.6 million in state mortgage-recording-tax exemptions, for a total of $1.463 billion. The summary does not treat the project as receiving a property-tax abatement, because the proposed PILOT and host-community payments are projected to exceed estimated full-assessment property taxes over time. (gcedc.com)
The Project’s Claimed Benefits
In the sponsor’s combined state-and-local benefit/cost analysis, which covers five years of construction and 30 years of operations, the proposed STAMP data center is presented as generating about $2.217 billion in discounted benefits against about $1.405 billion in discounted public costs, for an overall ratio of roughly 1.58 to 1. The claimed benefits include about 6,000 construction jobs over five years, with roughly $505 million in construction payroll, more than $564 million in operating payroll over 30 years, and about $278 million in additional payroll generated indirectly during construction and operation.
The report also counts more than $283.9 million in PILOT and host-community payments, about $18 million annually in combined state and local sales-tax revenue from electricity purchases, a little more than $5 million in fire-district fees, a $1.9 million zoning-related payment, $268 million to complete electrical infrastructure at STAMP, and $146 million in project-fee funding for future economic-development projects. Taken together, these assumptions produce the sponsor’s claim that the project would yield a positive combined return to state and local governments and residents.
The sponsor’s benefit/cost analysis relies on an unusually long effective time horizon: five years of construction followed by thirty years of operating-period benefits. That length matters because much of the project’s positive results depend on benefits assumed to arrive far in the future rather than during the period when public support is actually being provided. For a large, technology-intensive data center, a 35-year elapsed horizon is too long to serve as the main decision metric because it requires confidence in decades of future operating conditions, power costs, tax treatment, staffing levels, and continued competitiveness that cannot be known with much precision today. The longer the horizon, the more the reported ratio reflects assumptions rather than observable facts.
In this case, the long time period also helps the analysis by allowing decades of payroll, PILOT payments, and electricity-tax revenue to accumulate after the public subsidies are effectively delivered in the early years. A much shorter horizon — closer to the period in which public costs are incurred and project performance can be reasonably evaluated — would provide a more credible test of whether the project is likely to deliver a real return to New York residents.
There are several additional problems in the sponsor’s analysis. Most importantly, the sponsor gives full or near-full benefit credit to categories that do not clearly represent independent gains to New York residents, including $268 million in electrical infrastructure spending and $146 million in IDA fee-funded future projects, even though those items may primarily serve the project itself or are not directly tied to identifiable resident benefits.
The assumption that projected labor impacts can be treated largely as net new local gain is also unrealistic. The sponsor’s own report says 60% of the construction workforce is expected to come from outside the surrounding 14-county region. Yet, the analysis does not meaningfully discount the benefit of construction payroll for leakage or displacement — that is, for the likelihood that some workers will be imported or drawn away from other employers rather than representing new income to residents.
Finally, the analysis should not treat project-serving infrastructure as a public benefit without showing that it would provide substantial independent value to residents even if the project did not go forward. These assumptions materially overstate the project’s net benefit.
A revised benefit/cost analysis using more realistic assumptions found that the project’s economic case is much weaker than the sponsor’s headline presentation suggests. Using a shorter, more decision-relevant 10-year horizon, counting direct and indirect job benefits, along with public revenues that clearly accrue to New York residents, and excluding project-serving infrastructure, I estimated a benefit/cost ratio of about 0.58 to 1. Considering displacement, the ratio is 0.36:1 after a moderate displacement adjustment.
When my analysis was extended to match the sponsor’s much longer time frame — five years of construction plus thirty years of operating-period benefits — the ratio improved to about 1.13 to 1 before displacement. Still, it falls to about 0.85-1 once displacement is recognized. In other words, the analysis suggests that the project approaches break-even only if it is given a very long horizon and if most projected labor benefits are treated as net new gains; under more conservative assumptions, the return to New York residents remains below the project’s public cost.
Taken together, these issues suggest that the sponsor’s analysis materially overstates the project’s likely return to New York residents. Its positive result depends on a very long and highly uncertain time horizon, broad benefit categories that include items of doubtful independent public value, and labor assumptions that give too much credit to wages that may leak out of the region or displace other activity. When those assumptions are tightened, the project’s benefit/cost ratio falls well below the sponsor’s headline claim and, in the more realistic cases, remains below break-even. The basic conclusion is that the project’s economic case is much weaker and far more uncertain than the sponsor’s analysis implies.
Lost Opportunity
The project also raises important opportunity-cost questions. STAMP is a scarce public asset: a large, publicly assembled industrial site with major infrastructure commitments and limited electric capacity. If a very large share of that capacity and land is devoted to a data center with relatively modest permanent employment, the region may be giving up the chance to attract other users that could produce more jobs, broader supply-chain activity, or a stronger return per dollar of public support. The issue is not simply whether the data center creates benefits, but whether it is the best available use of the site, the infrastructure, and the subsidy package.
That concern is heightened here because the project is expected to require about 500 MW of power at a site whose electric plan is built around roughly 600 MW of substation capacity, meaning one tenant could consume most of the site’s near-term electric potential. In that setting, the true cost of the project is not just the tax benefits it receives, but also the value of alternative investments, tenants, and public uses that may be foreclosed or delayed if STAMP becomes a power-intensive, low-employment campus primarily.
Conclusions
Taken as a whole, the evidence suggests that the proposed STAMP data center is a weak case for large-scale public support. The sponsor’s analysis presents the project as producing a positive return. Still, that conclusion depends on a very long and highly uncertain time horizon, broad benefit categories that include items of doubtful independent public value, and labor assumptions that appear to overstate the share of benefits likely to remain with New York residents. When those assumptions are tightened — by using a shorter, more decision-relevant horizon, excluding project-serving infrastructure, and accounting for leakage and displacement in the labor market — the benefit/cost picture becomes much less favorable. In the revised analysis, the project does not deliver a clearly positive return over 10 years, and even on the sponsor’s much longer horizon, it falls below break-even after a moderate displacement adjustment.
That does not mean the project would have no value. It would generate construction activity, some permanent employment, real public revenues through PILOT and host-community payments, and some additional tax receipts. But the central policy question is not whether the project creates any benefit at all. It is whether the benefits are large enough, certain enough, and timely enough to justify roughly $1.46 billion in requested state and local tax subsidies and the commitment of scarce STAMP land, electric capacity, water, wastewater, and public attention to a single very large, power-intensive use. On that question, the case is much weaker. The project’s positive result appears to depend on counting decades of future benefits that may or may not materialize, while assigning full value to infrastructure and fee-related items that may primarily serve the project itself.
The opportunity-cost issue reinforces that conclusion. STAMP is a scarce public asset, and this proposal would consume an extraordinary share of its electric capacity for a facility expected to create relatively few permanent jobs compared with its scale. If the site and its infrastructure are instead available for a more employment-intensive or broadly rooted industrial use, the true cost of approving the project could be larger than the sponsor’s analysis suggests.
At the same time, the project’s size raises broader concerns about electric-system impacts, especially because very large new loads can increase pressure on generation, transmission, and prices even if regulators attempt to assign project-specific upgrade costs directly to the developer. New York’s own PSC has recognized that such projects can impose substantial costs on the electric system and has opened a proceeding to ensure that large loads pay their fair share.
The bottom line is that the sponsor has shown that the project could generate substantial economic activity. Still, it has not made a compelling case that it is a sound public investment on terms that are realistic, timely, and clearly beneficial to New York residents. A project that requires massive up-front subsidy, depends on a highly speculative long-run operating life to approach a favorable return, and offers limited permanent employment relative to its scale should be judged cautiously.
If public support is to be justified, the burden should be on the sponsor to show a credible positive return within a much shorter horizon, to distinguish clearly between project-serving and public-serving infrastructure, and to demonstrate that ratepayers and taxpayers will not bear costs that outweigh the project’s promised benefits.