By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel consensus is that the 10GW data-center project is risky due to its high cost, reliance on multiple partnerships, and potential political/regulatory challenges. The 'Ratepayer Protection Pledge' is a significant concern as it could lead to large, politically charged bills for ratepayers and invite legislative pushback.
Risk: The political/regulatory tail risk of the 'Ratepayer Protection Pledge' and the potential for state-level political veto power to derail the project.
Opportunity: If successfully executed, the project could provide Nvidia with a significant, long-term hardware demand and lock in OpenAI's compute capacity.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
OpenAI is in negotiations to lease a proposed 10-gigawatt data center campus on federal land in southern Ohio, with Nvidia (NASDAQ: NVDA) potentially backing the project financially, The Information reported on Wednesday, citing two people with direct knowledge of the discussions.
The campus, which could cost at least $500 billion to build at current prices for chips, labor, and power, would be developed by SB Energy, a SoftBank-backed power developer, Reuters further confirmed. The reported deal is structured as a 20-year lease, with OpenAI controlling the equipment at the facility and payments beginning only once operations start. The first phase is expected to deliver 800 megawatts of capacity in 2028.
Nvidia would supply hardware for the site and could provide a financial guarantee for OpenAI’s lease and SB Energy’s financing, according to the report. Reuters said it could not independently verify the details.
The site is tied to the Portsmouth Site in Pike County, Ohio, a former uranium enrichment facility that produced weapons-grade material during the Cold War before ceasing operations in 2001. The U.S. Department of Energy announced a public-private partnership with SoftBank and AEP Ohio in March 2026 to redevelop the land, including 10 gigawatts of new power generation and $4.2 billion in transmission upgrades.
Of the planned generation, 9.2 gigawatts would come from natural gas plants funded by $33.3 billion in Japanese capital tied to the U.S.-Japan Strategic Trade and Investment Agreement.
“Thanks to President Trump, the U.S. government is leveraging its assets, like our federal lands, to add power generation, create jobs, and ensure the United States wins the AI race,” U.S. Energy Secretary Chris Wright said in the March announcement.
## Ratepayer costs and grid buildout
The infrastructure costs are meant to be covered by the companies involved rather than passed on to households, under the White House’s Ratepayer Protection Pledge. OpenAI, Oracle (NYSE: ORCL), and xAI were among the companies that signed the pledge, which requires firms to “build, bring, or buy new generation resources and cover the cost of all power delivery infrastructure upgrades required for their data centers.”
“If it were not for the partnership between all parties, the Administration, SoftBank and our team, this type of investment would not be possible,” AEP Chairman and CEO Bill Fehrman said. “This partnership unlocks billions of dollars of electric transmission infrastructure, all without increasing customer rates.”
Four leading AI models discuss this article
"The stated 10 GW, $500B-cost data-center campus on federal land is likely not feasible in the near term given credible grid, capital, and regulatory risks, making the immediate read overly optimistic."
The article presents a jaw-dropping 10 GW campus financed via a web of partnerships, with 800 MW live by 2028 and a $500B price tag. Key red flags: costs seem implausibly high for a data-center build combined with generation assets; timeline hinges on federal land use, massive transmission upgrades, and long-term guarantees that shift risk to ratepayers. Verifiability is weak (Reuters and others couldn't confirm details), and quotes (e.g., Trump-era framing) strain credibility. The plan relies on OpenAI/ NVDA backing and Japanese capital to justify outsized capital outlays—exposure to policy shifts, geopolitical risk, and grid-build uncertainty could derail the thesis well before scale-up.
The numbers might be overstated or misinterpreted; a wall of capital and policy support could still surface, but the project may never reach the claimed scale or timing.
"OpenAI's transition from a software firm to an infrastructure-backed power consumer creates a new, defensible moat for Nvidia's hardware dominance."
This 10GW project is a massive bet on vertical integration, effectively turning OpenAI into a utility-scale power operator. By bypassing the public grid via the 'Ratepayer Protection Pledge,' OpenAI avoids the regulatory backlash of surging residential electricity prices. However, the $500 billion price tag is staggering—nearly double the current market cap of many S&P 500 components. If this succeeds, NVDA becomes the primary beneficiary of a sovereign-scale compute buildout. The risk is that this model creates a 'compute silo' that is geographically inflexible and highly vulnerable to natural gas price volatility, which accounts for over 90% of the planned generation capacity.
The sheer scale of this project creates massive execution risk; if natural gas prices spike or if the 2028 timeline slips, OpenAI could be left with billions in stranded lease liabilities and a massive stranded asset in rural Ohio.
"This deal is bullish for Nvidia's near-term revenue visibility but bearish for returns if AI capex intensity normalizes or efficiency gains reduce per-unit compute demand by 2027-2028."
This is a structurally clever deal that solves OpenAI's immediate compute bottleneck while offloading execution risk. The 20-year lease with payments deferred until 2028 operations start is a financing masterstroke—OpenAI gets 10 GW of committed capacity without balance-sheet strain, while Nvidia's implicit guarantee anchors the entire $500B+ project. The ratepayer protection pledge is politically durable. However, the article omits critical details: What's the actual power cost per MWh? What happens if chip costs deflate 30% by 2028 (plausible)? Can SB Energy actually execute a 9.2 GW natural gas build on schedule? The Portsmouth site's Cold War legacy adds permitting risk. Most importantly: this assumes AI capex intensity stays parabolic. If model efficiency gains or inference optimization reduce demand growth, this becomes stranded capacity.
A $500B+ facility betting on 2028 demand assumes AI scaling laws hold and no major efficiency breakthrough emerges in the next 3 years—a risky assumption given rapid progress in model compression and inference optimization already underway.
"The Ohio project highlights Nvidia's hardware centrality but layers multi-year political, financing, and infrastructure risks that the headline underplays."
The reported 10 GW Ohio campus would represent an unprecedented scale of AI infrastructure spend, locking in long-term Nvidia hardware demand starting 2028 while shifting power and transmission costs away from ratepayers via the DOE-SoftBank partnership. Yet the $500 billion build cost, 20-year lease structure, and dependence on $33.3 billion in Japanese natural-gas financing plus political continuity introduce execution and timing risks that dwarf typical data-center projects. First-phase 800 MW delivery remains years away, and any slippage in generation or transmission upgrades would directly hit Nvidia's visibility into sustained high-margin chip sales.
Even if negotiations succeed, the project's reliance on new gas plants and federal land redevelopment faces permitting and environmental delays that could push meaningful capacity well past 2028, muting any near-term re-rating for NVDA.
"The Ratepayer Protection Pledge adds political/regulatory tail risk that could derail financing if demand underperforms, not just commodity or execution risks."
Gemini's "compute silo" risk is real, but the far bigger, underdiscussed flaw is the political/regulatory tail risk of the Ratepayer Protection Pledge. If 2028 demand underperforms or grid upgrades lag, ratepayers could face large, politically charged bills, inviting legislative pushback and refinancing pressures. That fragility isn't captured in a simple 'success would benefit NVDA' thesis and could break the financing even before 2030.
"The Ratepayer Protection Pledge creates a systemic regulatory arbitrage risk that shifts failure costs to the federal government rather than local utilities."
Claude, you’re missing the second-order effect of the 'Ratepayer Protection Pledge': it’s a regulatory Trojan horse. By shielding local ratepayers, OpenAI effectively creates a private enclave that bypasses PJM Interconnection’s standard market clearing. This creates a massive 'regulatory arbitrage' risk. If the project fails, the federal government—not local utilities—will be the bag-holder. This isn't just a compute silo; it’s a potential sovereign-level bailout scenario if the 2028 operational targets evaporate.
"The Ratepayer Protection Pledge fails not via federal bailout but via state legislative revolt if rates spike."
Gemini's 'sovereign bailout' framing overstates federal exposure. The DOE isn't guaranteeing the $500B—SoftBank and OpenAI are. If 2028 misses, ratepayers face refinancing pressure, not Treasury backstop. The real tail risk is political: Ohio voters reject rate hikes, legislature blocks the pledge renewal, and OpenAI walks. That's a financing death spiral, not a bailout scenario. Nobody's flagged the state-level political veto power here.
"Ohio veto power would stall federal transmission upgrades and Nvidia demand visibility before refinancing becomes the issue."
Claude flags Ohio's legislative veto as a potential death spiral but misses how it directly threatens the DOE-SoftBank transmission upgrades tied to federal land at Portsmouth. State blockage would stall the 800 MW 2028 phase regardless of SoftBank's capex absorption, cutting Nvidia's visibility into sustained demand even before any refinancing hits. This state-federal linkage compounds the permitting risks already noted in the Cold War site.
The panel consensus is that the 10GW data-center project is risky due to its high cost, reliance on multiple partnerships, and potential political/regulatory challenges. The 'Ratepayer Protection Pledge' is a significant concern as it could lead to large, politically charged bills for ratepayers and invite legislative pushback.
If successfully executed, the project could provide Nvidia with a significant, long-term hardware demand and lock in OpenAI's compute capacity.
The political/regulatory tail risk of the 'Ratepayer Protection Pledge' and the potential for state-level political veto power to derail the project.