Watermark.
Fort Wayne GCP
Switch site 35 sites · 10 basins2 states Group by
OH Ohio 34
BOSC BOSC Lima Ottawa River · Lima, OH Building Under construction #1261 URB Urbana Mad River · Great Miami Live Investigating #1263 DEF Defiance Maumee mainstem Queued Investigating #1264 FIN Findlay Blanchard River Queued Investigating #1265 TOL Toledo Lucas Co WRRF Queued Investigating #1266 VWT Van Wert Town Creek · Little Auglaize Queued Investigating #1267 BRY Bryan Prairie Creek · Tiffin River Queued Investigating #1268 OTW Ottawa Blanchard River (lower) Queued Investigating #1269 SPR Springfield Mad River · Great Miami Queued Investigating #1270 XEN Xenia Little Miami Queued Investigating #1271 WPA Dayton · WPAFB Mad River · Great Miami Queued Investigating #1272 HAM Hamilton · Middletown Great Miami (lower) Queued Investigating #1273 TRP Troy · Piqua Great Miami (upper) Queued Investigating #1274 SID Sidney Great Miami · headwaters Queued Investigating #1275 GRV Greenville · Darke Co Stillwater · basin divide Queued Investigating #1276 WIL Wilmington Todd Fork · Little Miami Queued Investigating #1277 WUN West Union · Adams Co Ohio Brush Creek · Ohio River Queued Investigating #1278 NAL New Albany · Licking Scioto ↔ Muskingum divide Tracking Investigating #1279 COL Columbus Scioto · Olentangy Tracking Investigating #1280 CSH Coshocton Tuscarawas + Walhonding Tracking Investigating #1281 PIK Piketon Scioto River · PORTS Tracking Investigating #1282 SAN Sandusky · Perkins Twp Sandusky Bay · Lake Erie Tracking Investigating #1283 NWK Newark Licking River Tracking Investigating #1284 ZAN Zanesville Muskingum mainstem Tracking Investigating #1285 FRE Fremont · Clyde Lower Sandusky Tracking Investigating #1286 TIF Tiffin Sandusky (mid) Tracking Investigating #1287 BUC Bucyrus Sandusky headwaters Tracking Investigating #1288 CLE Cleveland Lower Cuyahoga Tracking Investigating #1289 AKR Akron Upper Cuyahoga · CVNP Tracking Investigating #1290 LRD Lordstown · Warren Upper Mahoning Tracking Investigating #1291 YNG Youngstown Mahoning mainstem Tracking Investigating #1292 LAN Lancaster Upper Hocking Tracking Investigating #1293 ATH Athens Lower Hocking Tracking Investigating #1294 LOG Logan Hocking Hills Tracking Investigating #1295
MAU Maumee Basin 8 sites
2MI The Two Miamis 9 sites
SE Southeastern Basins 10 sites
NE Northeast Basins 8 sites
Each site is a section of the Watermark platform — pivoted by basinstate. Live & building sites open; queued and tracking sites route to their coming-soon / watch page.
Switch site 35 sites · 10 basins2 states 35 open Group by

Each site is one Watermark investigation. The page color says what opens; the campus line is the build on the ground — two different clocks.

OH Ohio 34
BOSC BOSC Lima Ottawa River · Lima, OH Draft Under construction #1261 URB Urbana Mad River · Great Miami Open Investigating #1263 DEF Defiance Maumee mainstem Queued Investigating #1264 FIN Findlay Blanchard River Queued Investigating #1265 TOL Toledo Lucas Co WRRF Queued Investigating #1266 VWT Van Wert Town Creek · Little Auglaize Queued Investigating #1267 BRY Bryan Prairie Creek · Tiffin River Queued Investigating #1268 OTW Ottawa Blanchard River (lower) Queued Investigating #1269 SPR Springfield Mad River · Great Miami Queued Investigating #1270 XEN Xenia Little Miami Queued Investigating #1271 WPA Dayton · WPAFB Mad River · Great Miami Queued Investigating #1272 HAM Hamilton · Middletown Great Miami (lower) Queued Investigating #1273 TRP Troy · Piqua Great Miami (upper) Queued Investigating #1274 SID Sidney Great Miami · headwaters Queued Investigating #1275 GRV Greenville · Darke Co Stillwater · basin divide Queued Investigating #1276 WIL Wilmington Todd Fork · Little Miami Queued Investigating #1277 WUN West Union · Adams Co Ohio Brush Creek · Ohio River Queued Investigating #1278 NAL New Albany · Licking Scioto ↔ Muskingum divide Watching Investigating #1279 COL Columbus Scioto · Olentangy Watching Investigating #1280 CSH Coshocton Tuscarawas + Walhonding Watching Investigating #1281 PIK Piketon Scioto River · PORTS Watching Investigating #1282 SAN Sandusky · Perkins Twp Sandusky Bay · Lake Erie Watching Investigating #1283 NWK Newark Licking River Watching Investigating #1284 ZAN Zanesville Muskingum mainstem Watching Investigating #1285 FRE Fremont · Clyde Lower Sandusky Watching Investigating #1286 TIF Tiffin Sandusky (mid) Watching Investigating #1287 BUC Bucyrus Sandusky headwaters Watching Investigating #1288 CLE Cleveland Lower Cuyahoga Watching Investigating #1289 AKR Akron Upper Cuyahoga · CVNP Watching Investigating #1290 LRD Lordstown · Warren Upper Mahoning Watching Investigating #1291 YNG Youngstown Mahoning mainstem Watching Investigating #1292 LAN Lancaster Upper Hocking Watching Investigating #1293 ATH Athens Lower Hocking Watching Investigating #1294 LOG Logan Hocking Hills Watching Investigating #1295
MAU Maumee Basin 8 sites
2MI The Two Miamis 9 sites
SE Southeastern Basins 10 sites
NE Northeast Basins 8 sites
The page — what opens on Watermark Open · Draftfull record Queued · Watchingpreview page Restrictedrequest access
The campus — the build on the ground Investigating Confirmed Under construction Operational
Live

← The record · Legal history

Select Committee on Data Centers (2026)

Data Appendix — Cloud Pricing, AI Hardware, and Facility Resource Economics

Companion to the Response Fact Sheet — Ohio Select Committee on Data Centers

Witness: Cory Parent — interested party (cloud infrastructure engineer)

Dated: Reference figures as of May 2026 · prepared but not submitted to the Committee.

Source: bosc-data-appendix-2026-06-01.pdf (4 pp., text layer)

Faithful reproduction of the appendix text. Figures are industry reference ranges with the witness’s own cited sources (footnotes 1–7) — representative magnitudes, not facility-specific values for any single data center. Three points the figures establish: a government cloud carries a real, recurring price premium; AI-class hardware is an order of magnitude more power-dense and costly than the equipment a sales-tax-exemption forecast was built around; and a single large facility’s power and water footprint is a community-scale resource commitment.


1. The government-cloud premium

Government and sovereign cloud environments cost more than equivalent commercial cloud — physical isolation, U.S.-persons staffing, and compliance overhead rather than added compute. The premium is recurring: it applies to every hour and every gigabyte, for the life of the workload.

Representative commercial vs. GovCloud pricing (AWS)

Service / exampleCommercial AWSAWS GovCloudPremium
EC2 t3.medium (on-demand, per hour)$0.0416$0.0520~25%
S3 standard storage (per GB-month)$0.0240$0.0288~20%
EC2 compute, general rangebaseline+20–25%20–25%
Overall GovCloud vs. commercialbaseline+20–30%20–30% ¹

Sources: CapLinked GovCloud pricing analysis (2026); AWS EC2 on-demand pricing; Matlock LLC defense-cloud guide (5–25% application-infrastructure range). Representative on-demand rates; reserved/committed-use pricing differs.

Not unique to one vendor: Boston Consulting Group documents sovereign-cloud premiums of up to 30% industry-wide; Microsoft, Google, and Oracle all operate higher-cost government clouds (Azure Government, Google Distributed Cloud / Assured Workloads). ²

2. AI-class hardware: cost and power density

The sales-tax exemption is scored against a forecast of equipment purchases. AI-class hardware changes both terms: each unit costs far more than conventional servers, and it is replaced on a short cycle. A single high-end GPU server can exceed half a million dollars, and per-rack power is ~an order of magnitude above a conventional rack.

GPU and server hardware costs (purchase)

ItemApprox. purchase costNotes
NVIDIA H100 (single GPU)$25,000–$40,000Hopper generation
NVIDIA B200 (single GPU)$30,000–$50,000Blackwell generation
8-GPU H100 server$200,000–$450,000Complete system
DGX B200 (8-GPU server)~$515,000Complete system
DGX B300 (8-GPU server)$300,000–$350,000Blackwell Ultra (2026)

Sources: IntuitionLabs NVIDIA GPU pricing guide (Apr 2026); GMI Cloud H100 analysis (2026); aitooldiscovery Blackwell guide (2026). Purchase implies ~30–40% of hardware cost annually in operating cost and 6–12 month procurement. ³

Representative cloud rental rates (per GPU-hour, on-demand)

GPUHyperscaler on-demandNeo-cloud / spot
H100$4–$8 / hr~$2 / hr (PCIe); ~$1 spot
H200$3.72–$10.60 / hrfrom ~$3.80 / hr
B200~$14–$18 / hr~$2.12 / hr spot

Sources: Spheron GPU pricing comparison (May 2026); GMI Cloud (2026); Jarvislabs H200 guide (2026). Reserved pricing runs 20–40% below on-demand.

Power density: conventional vs. AI racks

Rack typePower per rackCooling
Traditional enterprise rack5–15 kWAir
H100-class AI rack40–50 kWAir / liquid
B200-class AI rack60+ kWLiquid
GB200 NVL72 rack-scale system120–140 kWLiquid (mandatory)

Sources: NVIDIA HGX data-center physical-requirements guidance, via IntuitionLabs (2025–2026). A single H100 GPU draws ~700 W; an 8-GPU H100 server ~5.6 kW before cooling/networking. Industry projections reach 250–900 kW per rack by 2027.

3. Facility power and water economics

At the facility level, the figures translate into a community-scale resource commitment. Power capacity is the defining metric — the same 25-MW line used in Ohio’s utility tariff and the proposed amendment — and cooling water, often withheld as proprietary, can reach millions of gallons a day depending on the cooling method.

Facility power capacity (typical ranges)

Facility scalePower capacity
Mid-size enterprise data centerunder 5 MW
Smaller hyperscale site10–30 MW
Typical hyperscale facility20–40 MW (new builds 50–100 MW)
Large hyperscale campus100 MW – 1 GW
Ohio tariff / amendment threshold (reference)25 MW

Sources: C&C Technology Group hyperscale power analysis (2025); TechPlusTrends AI data-center power guide (2026). A 50–100 MW facility can use as much electricity in a year as 400,000+ EVs.

Water use and efficiency (cooling)

Metric / facilityValueBasis
Industry-average water efficiency (WUE)~1.8–1.9 L/kWhThe Green Grid
Mid-size center, open-loop cooling~110M gal/yrEESI
Large hyperscale, evaporative coolingup to 5M gal/dayEESI / Vertiv
Efficient (closed-loop / chip cooling)~0.2–0.3 L/kWhMeta; Microsoft
U.S. data-center water use, 2023~17B gal (could double by 2028)TechTarget

Sources: EESI; Vertiv WUE analysis (2025); dgtlinfra / Meta and Microsoft sustainability disclosures (2024–2026); TechTarget. Cooling-water blowdown discharge is typically 20–40% of cooling-system water use — the wastewater dimension relevant to municipal treatment capacity.

Two caveats keep these honest. Water efficiency varies enormously by cooling method and climate — the same operator can report 0.2 L/kWh at one site and over 1.5 L/kWh in an arid region — which is why facility-level disclosure, not a portfolio average, is what a host community needs. And power capacity, not square footage, determines a facility’s grid and fiscal footprint, which is why a capacity-based definition is the workable one.


Sources

  1. Government-cloud premium. AWS GovCloud compute ~20–25% above commercial AWS; GovCloud overall 20–30% higher (EC2 t3.medium $0.0416 vs $0.0520/hr ~25%; S3 $0.024 vs $0.0288/GB-month ~20%). Defense-cloud guide cites 5–25% application-infrastructure range. CapLinked, “GovCloud Pricing in 2026”; AWS EC2 on-demand pricing; Matlock LLC, 2026.
  2. Sovereign-cloud premium, industry-wide. BCG documents sovereign-cloud premiums up to 30% (isolation, U.S.-persons staffing, compliance). Azure Government, Google Distributed Cloud / Assured Workloads, Oracle comparable. BCG, “Cloud Cover,” Aug 2025; vendor docs 2025–2026.
  3. AI GPU/server purchase costs. H100 ~$25–40k; B200 ~$30–50k; 8-GPU H100 servers ~$200–450k; DGX B200 ~$515k; DGX B300 (shipped Jan 2026) ~$300–350k. ~30–40% of hardware cost annually in opex; 6–12 month procurement. IntuitionLabs; GMI Cloud; aitooldiscovery, 2026.
  4. AI GPU cloud rental (on-demand, per GPU-hour). H100 ~$4–8 (hyperscaler), ~$2 PCIe / ~$1 spot; H200 ~$3.72–10.60; B200 ~$14–18, ~$2.12 spot. Reserved 20–40% below on-demand. Spheron, May 2026; GMI Cloud; Jarvislabs, 2026.
  5. Rack power density and cooling. Enterprise ~5–15 kW; H100-class 40–50 kW; B200-class 60+ kW; GB200 NVL72 120–140 kW (liquid mandatory). Single H100 ~700 W; 8-GPU H100 server ~5.6 kW pre-cooling. Projections 250–900 kW/rack by 2027. NVIDIA HGX guidance via IntuitionLabs, 2025–2026.
  6. Facility power capacity. Mid-size <5 MW; smaller hyperscale 10–30 MW; typical 20–40 MW (new builds 50–100 MW); large campuses >100 MW, AI “gigafactories” ~1 GW. 50–100 MW ≈ 400,000+ EVs/yr. The 25 MW reference matches the AEP Ohio tariff and proposed amendment thresholds. C&C Technology Group, 2025; TechPlusTrends, 2026.
  7. Facility water use and efficiency. WUE ~1.8–1.9 L/kWh; mid-size open-loop ~110M gal/yr; large evaporative-cooled up to ~5M gal/day; efficient closed-loop / chip-cooled ~0.2–0.3 L/kWh. U.S. data centers ~17B gal in 2023, projected to ~double by 2028; cooling blowdown ~20–40% of cooling-system water use. EESI; Vertiv; dgtlinfra; Meta/Microsoft disclosures; TechTarget, 2024–2026.