Feasibility File · SoftBank

PORTS Technology Campus

Piketon · OH

Weak but not impossible. Real federal land, real AEP context, real DOE / U.S.–Japan sponsorship — and a 9.2 GW gas plant whose proof package has not entered the public record.

Verdict

3.1/ 10

Bearing Labs feasibility score

Treat PORTS as a weak but not impossible project. The public record supports a serious development effort at Piketon; it does not support the announced 9.2 GW / 10 GW buildout as deliverable on the stated schedule. Prove an 800 MW power path before crediting the 10 GW campus, align the 2028 data-center claim with actual AEP/PJM energization evidence, publish a cooling and water balance before claiming environmental feasibility, isolate consumer cost exposure through a project-specific order or contract, sequence DOE remediation, NEPA, Ohio EPA, ODNR, OPSB, PJM, and local approvals in one public critical path, and treat the January 2029 federal transition as a real durability deadline. Score should remain 3.1 until named evidence is public: executed turbine/EPC agreements, firm gas transportation and pipeline filings, PJM/AEP interconnection and network-upgrade commitments, OPSB-approved transmission route and schedule, PORTS-specific ratepayer-protection order or contract, first-phase water and cooling permits, and DOE construction/remediation interface approval.

Where the story diverges

What the operator says

10 GW AI data-center campus on the former DOE Portsmouth Gaseous Diffusion Plant in Pike County, Ohio. Proposed $33.3B / 9.2 GW natural-gas generation package alongside $4.2B in new AEP transmission, with power beginning to flow in 2029.

Outgrowth of the U.S.–Japan trade and investment framework; SoftBank has described potential multi-decade investment up to $1.5T. Construction targeted to start in 2026, with an initial $10B / 800 MW phase online by 2028 and the broader project advancing meaningfully before January 2029.

What's actually there

The site is directionally attractive for a smaller, phased redevelopment: large former federal industrial land, unusually strong bulk-transmission proximity (345 kV and 765 kV class infrastructure nearby), a region seeking replacement economic activity, and U.S.–Japan political sponsorship. The claim mean of 3.1 is itself revealing: PORTS is not being held down by one narrow gap. The weakness is distributed across power, schedule, gas, water, permitting, cost, workforce, and political durability.

A 9.2 GW gas plant, ~1.2 Bcf/day of firm gas deliverability, PJM reliability review, AEP siting and construction, new water and discharge permits, DOE remediation interfaces, and a workforce ramp toward 35,000 construction workers all have to clear together. The controlling risk is firm power deliverability plus project-specific cost allocation: it caps the published score at 4.5 even before the broader weaknesses are counted. AEP's public transmission schedule pointing to 2029 power flow also undercuts any claim that 2028 power availability for the data-center phase is already solved.

Risk 01

Single most decision-relevant exposure

Firm power deliverability plus project-specific cost allocation

The controlling risk is whether SoftBank, AEP Ohio, and PJM can show firm 9.2 GW gas deliverability and customer-funded transmission and gas-system upgrades — with collateral, take-or-pay terms, exit fees, and stranded-cost protection — before approvals and construction lock in. Generic intent and a strong AEP tariff framework are not the same as a PORTS-specific cost-allocation order. Without that proof, the project sits at a capped 4.5 and the broader cluster of weak claims keeps the published score at 3.1.

Claim Scorecard

Every announced claim, scored.

Each claim assessed independently against supporting and contradicting evidence in the full report.

01
3.5/ 10
Power delivery and turbine supply (9.2 GW gas plant as scoped)
Meaningful but qualified
Sponsor seriousness, DOE land, and AEP involvement are real; turbines, EPC, air permits, gas, and construction have not entered the public proof record.
3.5/ 10
02
2.0/ 10
Timeline (2026 construction start · 800 MW by 2028 · before Jan 2029)
Meaningful but qualified
Weakest claim in the scorecard. AEP's 2029 power-flow path undercuts any 2028 first-phase commercial-operation claim absent extraordinary evidence.
2.0/ 10
03
3.0/ 10
Gas supply adequacy (~1.2 Bcf/day)
Preliminary
Statewide production does not prove plant-gate deliverability, pipeline capacity, firm transport, or price-risk allocation for a 9.2 GW fleet.
3.0/ 10
04
4.0/ 10
PJM interconnection and AEP transmission (~50-mile line)
Meaningful but qualified
Strongest infrastructure claim because location and AEP posture are real; still only a process claim. OPSB Case No. 26-0426 is an early marker, not approval.
4.0/ 10
05
3.0/ 10
Consumer cost protection
Meaningful but qualified
AEP Ohio data-center tariff framework is supportive in concept; no PORTS-specific cost-allocation order, special contract, or take-or-pay mechanism is public.
3.0/ 10
06
3.5/ 10
Economic and political durability
Meaningful but qualified
Federal and U.S.–Japan sponsorship is strong; economics, financing, gas exposure, and post-2029 political continuity are weakly proven.
3.5/ 10
07
3.0/ 10
Site suitability and water (former nuclear site)
Meaningful but qualified
Water source, cooling design, discharge permits, floodplain design, and DOE remediation-interface unresolved. EPA ID OH7890008983 cleanup runs into the early 2030s.
3.0/ 10
08
2.5/ 10
Workforce (region can supply 4k → 35k construction workers)
Preliminary
A 35,000-worker peak would be a multi-state mobilization event; no housing, transport, labor, or training plan is public.
2.5/ 10

Cross-Cutting Risk Factors

The interfaces that decide the project.

01
The project stacks megaproject risks instead of sequencing them
PORTS is simultaneously a data-center, gas-generation, transmission, potential pipeline, water / wastewater, federal land-reuse, and nuclear-site remediation-interface project. The schedule risk comes from interfaces: a delay in PJM, OPSB, DOE, Ohio EPA, ODNR, turbine procurement, gas transportation, or water-source design can block the others.
02
Power scale overwhelms ordinary data-center heuristics
A 9.2 GW gas plant is not a campus utility package — it is comparable to a large regional power buildout. Reference-class review shows delivery confidence rises only when sponsors can demonstrate long-lead equipment control, utility studies, cost-allocation mechanisms, and construction readiness. PORTS has ambition but not yet equivalent proof.
03
The 2028 / 2029 mismatch is a hard schedule warning
If AEP-linked power flow is expected in 2029, then a 2028 first-phase commercial-operation claim requires either earlier grid service, temporary bridge power, behind-the-meter generation, a smaller energized load, or a looser definition of 'online' — none of which is documented as the delivery basis.
04
Water and cooling choices determine environmental feasibility
The difference between dry, hybrid, wet, and liquid-cooling designs could move water demand by tens of millions of gallons per day. Without a cooling basis, the project cannot credibly claim water feasibility, wastewater feasibility, thermal-discharge feasibility, or public-health reassurance.
05
Cost protection is politically inseparable from feasibility
If PORTS can prove beneficiary-pays treatment, local benefits, and stranded-cost protection, community resistance may remain manageable. If consumers believe the project shifts risk to electricity and gas bills, the ratepayer issue can become a statewide political constraint, as seen in Northern Virginia and other data-center markets.
06
The former nuclear-site history cuts both ways
The PORTS site offers an unusually strong redevelopment story: productive reuse of federal land in a community that wants investment. It also creates a high-trust burden. DOE cleanup, groundwater plumes, demolition, radiological history, and waste-disposal interfaces are part of the delivery path, not background facts.

Project Profile

Reference data from the public record.

Project: SoftBank PORTS Technology Campus
Stage tag: Pre-permitting (Announced)
Site: Former Portsmouth Gaseous Diffusion Plant — DOE land in demolition / remediation, >3,700 historical acres
Developer: SB Energy (SoftBank subsidiary) with U.S. federal government / DOE
Data center: Proposed 10 GW AI campus · initial $10B / 800 MW phase
Power plant: Proposed $33.3B / 9.2 GW natural-gas generation package
Transmission: $4.2B AEP Ohio investment tied to the project; pathway for power to begin flowing in 2029
Gas demand: ~1.2 Bcf/day at 90% utilization vs. Ohio 2024 production of ~5.9 Bcf/day · ~438M MMBtu/yr
Timeline: Construction targeted to start 2026; first 800 MW phase by 2028 / before Jan 2029
Financing frame: U.S.–Japan trade/investment framework; SoftBank multi-decade narrative up to $1.5T
Stakeholders: SB Energy · DOE · AEP Ohio · GE Vernova · PJM · OPSB · Ohio EPA · Pike County · Government of Japan
Scoring arithmetic: Claim mean 3.1 · controlling claim 3.0 · cap 4.5 (does not bind)
Evidence as of: May 29, 2026

Bearing Labs · Feasibility Brief