← The record · Legal history

Select Committee on Data Centers (2026)

Written Testimony — Ohio Select Committee on Data Centers

Beneficial-Ownership Disclosure in Ohio Data Center Development

Witness: Cory Parent — Allen County resident & cloud infrastructure engineer (interested party)

Dated: June 1, 2026 · Delivered orally: 2026-06-04 (see oral index)

Source: bosc-written-testimony-2026-06-01.pdf (8 pp.)

Faithful reproduction of the submitted text (the PDF carries a text layer). To Co-Chairs Holmes and Chavez and members of the Select Committee on Data Centers.

Good afternoon, Co-Chairs Holmes and Chavez, and members of the Select Committee on Data Centers. My name is Cory Parent. I am a resident of Allen County and a software and platform engineer who builds and operates cloud infrastructure for regulated industries — healthcare, financial services, and commercial real estate. I work daily with the platforms these facilities are built to serve and with the economics that govern how their capacity is bought, sold, and consumed.

I am here to make one argument, and to make it plainly: Ohio extends substantial public benefits to data center developers without requiring them to disclose who they are. That gap should be closed. I am not opposed to data centers — Ohio should welcome them, and the bills already before this Committee get a great deal right. But every other safeguard the State is building depends on a question Ohio law does not currently make anyone answer: who is the customer?

The Gap: Ohio Subsidizes Developers It Cannot Name

Ohio grants tax abatements through the Department of Development, commits public water and power, and in some cases finances infrastructure — for entities that need never disclose their ultimate ownership to the public bodies extending those benefits. A shell company can hold the deed, sign the development agreement, and receive the abatement while the principal behind it stays hidden from the community paying for it.

I can speak to this directly. In American Township, in my county, a data center representing roughly half a billion dollars in capital investment was negotiated for about fifteen months under a non-disclosure agreement with a Delaware shell company formed for the purpose. The developer’s identity was withheld from the public the entire time, and key engineering specifications were redacted from a public environmental permit as proprietary. The local officials who approved abatements and committed water did so in good faith, on incomplete information, because nothing in Ohio law required more. This is not a criticism of those officials. It is a description of a gap in the law — and it is fixable. [1]

A Common Definition the Whole Code Can Use

Before the State can disclose who is behind a facility, it has to agree on what the facility is. The only definition of “computer data center” in the Revised Code lives in the tax-exemption statute (R.C. 122.175(A)(2)) — written to administer that one program, tied to capital-investment and payroll minimums. [2] It is circular and program-bound: it says nothing about size, load, or function, and was never meant to govern zoning, water allocation, utility service, or environmental permitting. The proposed constitutional amendment shows the alternative instinct — it defines a data center by a measurable threshold, an aggregate or peak load above twenty-five megawatts. [3] Ohio should adopt one standard definition — capacity-based, technology-neutral, and able to aggregate facilities under common ownership — and use it across the Revised Code, so the abatement statute, the utility tariff, the water-reporting rule, and any size threshold all govern the same thing. Without it, every safeguard rests on a term each agency reads differently and a developer can structure around.

Why Disclosure Is the Foundational Fix

Nearly every other protection the State is pursuing quietly depends on disclosure. The cost-causation principle in the PUCO’s data center tariff and in House Bill 706 depends on knowing who the cost-causer is. The anti-speculation forecasting the Ohio Consumers’ Counsel urged depends on verifying who stands behind a load request — and PJM’s own load-forecasting reform now requires NDAs that at least permit sharing a load’s identity with the grid operator: an explicit acknowledgment that disclosure and reliability travel together. The 25-megawatt threshold in the proposed amendment cannot aggregate parcels to a common owner precisely because no disclosure requirement exists; hyperscale infrastructure is distributed by design and routes around a per-parcel ceiling as routine site planning. [4] And public trust erodes when a community cannot find out who it is being asked to subsidize — the petition’s own organizers used the words “cloaked in secrecy.” [5] Disclosure does not compete with these concerns; it is the common foundation underneath all of them.

What Ownership Hides: The Government-Cloud Premium

There is an economic dimension to ownership the public record cannot see, sharpest in a state with Ohio’s defense footprint. Government cloud (“GovCloud”) environments that host federal and defense workloads are not a labeling difference — they are a different cost structure and addressable market. Government cloud capacity runs roughly 20–30% above commercial rates because of physical isolation, U.S.-persons staffing, and compliance overhead. [6]

What drives the premium is the authorization level, which also dictates the physical facility. FedRAMP governs federal cloud work; the DoD impact levels form a ladder above it — IL4/IL5 for sensitive/national-security data, IL6 for classified up to SECRET. At the higher levels the rules require dedicated, physically isolated infrastructure and vetted U.S.-persons staffing; a facility built for IL5/IL6 is, by regulation, not the flexible shared capacity a commercial abatement forecast assumes. [7] This is not abstract in Ohio: the state hosts Wright-Patterson Air Force Base and an extensive defense supplier network, making high-authorization hosting a realistic end use here. [8] A facility serving that demand has a different margin, customer, and a far more durable revenue base — and when the State scores an abatement on commercial assumptions, the public may be underwriting a more valuable, more permanent enterprise than the one it was shown.

The authorization posture also governs who in Ohio may do business with the facility at all. A high-authorization enclave’s capacity is not sold to the open market; an IL5/IL6 environment cannot host a local hospital, a regional bank, or a county government as a tenant, because the isolation is the whole point. So a facility presented as shared digital infrastructure may, depending on end use, be capacity the community is structurally barred from using — built on local land, drawing local water and power, and closed to local enterprise. It also determines whether the facility seeds a local technology cluster or remains a sealed island: a government enclave’s supply chain is federal and out-of-region, its workforce constrained by clearance, its capacity closed to local ventures. The difference is a function of end use and ownership — exactly the facts disclosure would put on record.

The stakes are concrete in a two-state contrast. Ohio exempts data-center equipment from sales tax and levies no tangible personal property tax, and abates real-property improvements 75% for 15–30 years. Of thirteen data-center agreements approved through September 2024 (~$5.1B investment): 356 jobs, $31.6M annual payroll, against ~$281.9M state revenue loss — closer to $1M per job with local sales-tax losses included. [9] Virginia taxes data-center equipment as business personal property: Loudoun County draws ~38% of its general-fund revenue from data centers on ~4% of its commercial parcels (>$100M/yr), and has cut its residential property-tax rate every year for a decade. The same capital can fund schools and lower homeowner taxes, or occupy land and draw power while contributing almost nothing for a generation. The instrument that closes the gap is a community benefits agreement — and to date none has been negotiated for any Ohio data center, because NDA-governed negotiations present a community with terms already finalized. [10]

A Technical Caution on the Sales-Tax Exemption Forecast

The Data Center Tax Exemption is scored against a forecast of what the facility will buy. [11] That forecast is harder than it looks: the equipment does not sit still. Hyperscale facilities refresh servers on a 3–5 year cycle, and at AI-class densities per-rack hardware cost has risen an order of magnitude. [12] The exempted purchases are a recurring, escalating stream, not a one-time build-out; a forecast built on initial fit-out or conventional unit costs will understate the realized exemption substantially, and the gap compounds each refresh. The State should build refresh cycles and current unit costs into the forecast rather than scoring a single snapshot.

Where the Current Bills Get It Right — and Where They Stop Short

House Bill 706 (Reps. Thomas and Rader) is the strongest vehicle: it makes cost causation enforceable (long-term service agreements, minimum billing demand, exit fees, a bar on shifting data-center costs onto other ratepayers) and extends statewide a PUCO-confirmed tariff. [13] House Bill 646’s study commission is the right venue for questions that need study; the Electricity Forecast Integrity Act addresses speculation. These efforts are not insufficient — they are incomplete without disclosure, and disclosure makes each work better.

On HB 646: a study commission is only as good as the information it is fed. The parties with the most knowledge are the operators, who have every incentive to shape its conclusions; without an independent source of facility-level data, a commission is vulnerable to capture. On HB 706: keep its requirements anchored in retail service terms and align interconnection provisions with FERC’s PJM framework (Federal Power Act = FERC over wholesale cost creation; states over retail allocation) so it holds up if challenged.

Recommendations

1. Condition any public incentive, abatement, or infrastructure commitment on disclosure of the developer’s beneficial owner and controlling parent into the public record (cf. the federal Corporate Transparency Act framework and Ohio procurement). The foundational fix. [14]
2. Advance HB 706 as the core consumer-protection vehicle, anchored in retail service terms and aligned with FERC’s framework.
3. Use HB 646’s study commission to close the water-visibility gap Director Mertz identified — require data centers on public water systems to meter and report facility-level use — and recommend a load standard based on local grid headroom rather than a flat megawatt figure.
4. Require publicly accessible, near-real-time facility-level reporting of resource use (WUE/PUE) and disclosure of the facility’s federal authorization posture (any FedRAMP authorization or DoD impact-level accreditation, and any government-related equipment/capacity hosted on site). The data already exists in operators’ systems; what is missing is the requirement to publish per facility rather than rolled into a corporate aggregate.
5. Build on the forecasting and anti-speculation work already underway, so investment rests on demonstrated need and existing consumers are protected from stranded costs.
6. Treat each abatement as a negotiation, not a concession — with ownership and end use in hand, condition incentives on enforceable community benefits (permanent employment, resource use, local investment) of the kind no Ohio data-center agreement has yet secured.

Disclosure and Standing

The witness filed written comment in his community’s local zoning process before these circumstances became publicly known, establishing a good-faith record as a process participant. He discloses a recently incorporated govtech startup whose thesis is offering transparency and making enabling it a core design principle. The recommendations rest on structural facts that are a matter of public record, not on disclaiming that interest.

“Ohio can welcome this industry and protect the communities that host it. The framework is most of the way built. Requiring developers to disclose who they are completes it.”

Sources (14 footnotes, as submitted)

1. Ohio data center NDA secrecy / local-official disclosure — Reps. Brian Stewart & Adam Bird introduced HB 695 to prohibit certain local officials from signing economic-development NDAs. Ohio Society of CPAs / Ohio Capital Journal, Mar 2026.
2. Statutory definition — only definition of “computer data center” is R.C. 122.175(A)(2) (+ minimums at (A)(5)); program-specific, no size/load/function threshold. ORC; OAC 122:28-1-01.
3. Capacity-based definition — Ohio Prohibition of Construction of a Data Center Amendment (2026), proposed §36a, Art. II: aggregate monthly demand or peak load > 25 MW. Ballotpedia; Ohio Capital Journal, Apr 2026.
4. Aggregation gap — amendment sets 25 MW ceiling but no ownership-disclosure/parcel-aggregation; certified by Ohio Ballot Board (~413,000 signatures by Jul 1, 2026). Ballotpedia; OCJ 2026.
5. “Cloaked in secrecy” — ballot organizers/rural residents on NDA secrecy. Ohio Capital Journal; Cincinnati Enquirer, 2026.
6. GovCloud premium — AWS GovCloud ~20–30% above commercial (m5.large GovCloud US-East = 26% premium); BCG documents up to 30% for sovereign cloud (isolation, U.S.-persons, CLOUD Act). ProsperOps 2024; CapLinked 2026; BCG “Cloud Cover” Aug 2025.
7. FedRAMP/DoD impact levels — DoD CC SRG extends FedRAMP: IL4 (CUI), IL5 (higher CUI + NSS) need dedicated infra + U.S.-citizen staff; IL6 = classified up to SECRET, wholly dedicated, SIPRNet. DoD CC SRG.
8. Ohio defense footprint — Wright-Patterson AFB + supplier network; Google Distributed Cloud air-gapped appliance holds DoD IL5, MIL-STD-810H; Air Force Rapid Sustainment Office a named early customer; GDIT + Google Public Sector demoed at Exercise Mobility Guardian 2025. Google Cloud blog; Breaking Defense; Defense One; GDIT, 2024–25.
9. Ohio subsidy scale — Dept. of Development: 13 agreements through Sep 2024 (~$5.1B) → 356 jobs, $31.6M payroll vs ~$281.9M revenue loss (>$343M w/ local sales tax). Microsoft 2024: 90.9% of investment, 4.8% of new FTE jobs. Policy Matters Ohio Jan 2025; News 5 Cleveland 2026.
10. Virginia / Loudoun — taxes equipment as business personal property; ~38% of general fund from data centers on ~4% of commercial parcels (>$100M/yr); residential rate cut $1.145 (2016) → $0.805/$100 (2025). No Ohio CBA yet. Loudoun County; Policy Matters Ohio; WOUB, 2025–26.
11. Sales-and-use tax exemption — R.C. 122.175 (min $100M capital / 3 yrs, ≥$1.5M payroll), Tax Credit Authority; ~$554.9M in exemptions in 2024; no TPP tax. ORC; Tax Foundation; PMO; News 5, 2025–26.
12. Refresh cycles / AI rack costs — 3–5 yr refresh (3-yr TCO horizon for AI); traditional racks ~5–15 kW vs AI 40–120+ kW; GPU servers ~$200k–$450k (H100) to >$500k (B200). McKinsey; Dell’Oro; CHG-Meridian, 2025–26.
13. HB 706 (Thomas & Rader) — long-term service agreements, minimum billing demand, exit fees/collateral, bar on cost-shifting; extends the AEP Ohio tariff statewide. Ohio House; Statehouse News Bureau, Feb–Mar 2026.
14. Federal beneficial-ownership framework — Corporate Transparency Act (FinCEN); Mar 2025 interim final rule narrowed reporting to foreign-formed entities, exempting U.S.-formed. Concept remains established; State may require it as an incentive condition. FinCEN; Treasury, 2025.