$500 Billion, 20-Year Lease: OpenAI in Talks for 10GW Ohio Data Center, Nvidia Set to Act as Credit 'Backstop'

marsbitОпубліковано о 2026-06-10Востаннє оновлено о 2026-06-10

Анотація

OpenAI is in advanced negotiations with SB Energy, a SoftBank subsidiary, to lease a massive 10-gigawatt AI data center campus planned for a former uranium enrichment site in Pike County, Ohio. The proposed 20-year lease could involve total rental payments amounting to tens of billions of dollars, with the first 800-megawatt phase slated for 2028. A key structural innovation reported is that Nvidia is in discussions to act as a credit guarantor for OpenAI's lease payments and SB Energy's project financing. This marks a significant expansion of a financing model where major tech firms back AI startups' infrastructure commitments, following similar reported arrangements between Google and Anthropic. The project is part of a larger U.S.-Japan investment framework and involves SB Energy building 9.2 GW of gas-fired power generation. The total cost for the required IT hardware, primarily Nvidia chips, is estimated at around $350 billion, for which OpenAI is seeking separate financing. This deal represents a shift in OpenAI's strategy towards large-scale bilateral agreements for its own compute infrastructure, moving away from earlier joint venture plans like the shelved "Stargate" project. This massive infrastructure commitment coincides with OpenAI's confidential IPO filing. Analysts expect its long-term cloud and compute obligations, totaling over $665 billion, to be a major focus for SEC scrutiny and investor due diligence. SB Energy has also announced its own IPO plans, cre...

Author: Ada, Shenchao TechFlow

OpenAI's capital structure for computing infrastructure is being rewritten.

According to a June 10 report from The Information, OpenAI is in deep negotiations with SB Energy, a subsidiary of SoftBank Group, for a major lease. The subject is a planned AI data center campus on federal land in Pike County, southern Ohio, with a total planned capacity of 10 gigawatts. This campus was previously disclosed by the U.S. Department of Energy (DOE) and the Department of Commerce (DOC) in a "public-private partnership" format, to be built on the site of the former Portsmouth uranium enrichment plant, which produced weapons-grade uranium during the Cold War and shut down in 2001.

Under the terms disclosed by The Information, OpenAI will sign a 20-year lease with SB Energy, during which OpenAI will control the use of the equipment within the campus and will be obligated to pay rent after the project becomes operational. The cumulative rent over the lease period is expected to reach tens of billions of dollars. The first phase, with 800 megawatts of capacity, is expected to be operational by 2028.

Nvidia Steps In as Credit Guarantor, Structural Financing Model Begins to Spread

The core, and previously undisclosed, design of the deal lies in Nvidia's role. According to The Information, Nvidia is in discussions to use its own balance sheet to provide credit guarantees for OpenAI's lease payments and future project financing for SB Energy. This means Nvidia would, for the first time, act as a "financial guarantor" for an infrastructure project of this scale. Its previous collaborations with OpenAI have primarily involved equity investments and sale-leaseback financing structures.

This model is not an isolated case. According to a Bloomberg exclusive report on June 9, Google is essentially providing a backstop arrangement for Anthropic's TPU leasing obligations, which amount to approximately $35 billion, covering five data centers located in New York, Texas, Louisiana, and Indiana. That deal is supported by structured debt issued by a special purpose vehicle (SPV) led by Apollo and Blackstone, with Broadcom providing additional residual value guarantees on the senior tranche. SemiAnalysis describes this structure as a "hyperscaler-backed model," suggesting it is addressing the duration mismatch issue inherent in data centers, where "lease terms exceed 15 years while the payback period is about 8 years."

Nvidia's entry signals that this structure is spreading from cloud providers to the chip supplier side.

$500 Billion Campus Relies on U.S. DOE Land and Japan-U.S. Investment Agreement

The lessor, SB Energy, is an energy company established by SoftBank in 2019, with SoftBank holding a controlling stake. Ares Management and OpenAI are also shareholders. In January of this year, OpenAI and SoftBank Group each injected $500 million (totaling $1 billion) into SB Energy. Concurrently, OpenAI awarded SB Energy the contract to build the 1.2-gigawatt "Stargate" project in Milam County, Texas.

The project's governmental background is particularly crucial. According to a U.S. Department of Commerce disclosure in February, this SB Energy investment stems from the framework of the "Trade Agreement with Japan" reached during the Trump administration. Japan committed to investing $550 billion in the U.S., which includes SB Energy's $33 billion power plant project at the Portsmouth campus. In early May this year, Energy Secretary Chris Wright, Commerce Secretary Howard Lutnick (Note: This seems to be an error in the original text; likely meant a different official), and Interior Secretary Doug Burgum attended the project announcement ceremony in Piketon alongside representatives from SoftBank.

Under the plan, SB Energy will self-fund the construction of 9.2 gigawatts of natural gas power generation units (assets owned by the U.S. government, operated by SB Energy) and partner with AEP Ohio to invest $4.2 billion in grid upgrades. Upon completion, the campus's total 10-gigawatt capacity would be roughly equivalent to the output of 4.5 Hoover Dams, potentially making it the world's largest single data center campus.

$350 Billion Chip Shortfall to Fill, OpenAI Uses 'Piecemeal' Approach to Patch Infrastructure

Having just the campus facilities is far from sufficient. According to industry standards, chips and servers and other IT equipment account for about 70% of the total cost of an AI data center. For this project alone, OpenAI would need to raise approximately $350 billion to purchase Nvidia chips. The Information reports that OpenAI is already discussing how to finance this chip procurement separately.

This move reflects a shift in OpenAI's infrastructure strategy. Earlier this year, the $500 billion "Stargate" joint venture initiative launched by OpenAI, Oracle, and SoftBank was confirmed by the industry to have been effectively shelved mid-year. The FT reported that OpenAI has "effectively abandoned the joint venture model" and is turning to large-scale bilateral deals to advance its own computing power deployment. This 20-year lease with SB Energy is the largest single product of this new approach.

$665 Billion in Cloud Lease Obligations to Become Focus of IPO Prospectus Scrutiny

This Monday (June 8), OpenAI confirmed it had confidentially submitted its IPO prospectus (Form S-1) to the SEC. According to disclosures by CNBC and Reuters, the underwriting syndicate includes Goldman Sachs, Morgan Stanley, and JPMorgan Chase, with a target valuation potentially exceeding $1 trillion, aiming for a listing as early as September. Media outlets like Bloomberg, citing internal documents, predict OpenAI will still record a loss of approximately $14 billion in 2026 and may not achieve profitability before 2029.

Against this backdrop, the at least $665 billion in cloud leasing commitments OpenAI has signed over the past year with Microsoft (incremental $250 billion), Oracle ($300 billion over 5 years), and AWS ($100 billion total over 8 years, including an original $3.8 billion) will become a core focus for SEC review and institutional investor due diligence. OpenAI's CFO, Sarah Friar, admitted internally in April that if revenue growth falls short of expectations, the company might not be able to fulfill its future computing power contracts.

Meanwhile, SB Energy itself announced its plans for a U.S. IPO in May of this year. The underwriting syndicate includes JPMorgan Chase, Goldman Sachs, Morgan Stanley, Citigroup, and Mizuho, with a valuation potentially exceeding $50 billion, aiming for a listing as early as September. This creates a situation where OpenAI, SB Energy, and their upstream partner Nvidia will be "exposing each other's risk exposures in their respective prospectuses," adding another layer to the "circular financing" narrative.

Пов'язані питання

QWhat is the scale and key financial arrangement of the AI data center campus that OpenAI is reportedly negotiating in Pike County, Ohio?

AOpenAI is negotiating a 20-year lease with SB Energy for a planned 10-gigawatt AI data center campus on federal land in Pike County, Ohio. The total rent over the lease term is projected to reach hundreds of billions of dollars, with the first phase of 800 megawatts expected to be operational by 2028.

QWhat unprecedented financial role is NVIDIA reportedly playing in the Ohio data center deal involving OpenAI and SB Energy?

ANVIDIA is reportedly in discussions to act as a 'credit guarantor' for the deal, using its balance sheet to backstop OpenAI's lease payments and future project financing for SB Energy. This marks a first for NVIDIA in providing such financial guarantees for a large-scale infrastructure project of this nature.

QWhat broader industry financing pattern does the OpenAI-SB Energy deal, with NVIDIA's involvement, represent according to the article?

AThe article suggests the deal represents the spread of a 'hyperscaler-backed' structured financing model from cloud providers to chip suppliers. This model, also seen in deals like Google backing Anthropic's obligations, addresses the 'duration mismatch' between long-term leases (15+ years) and the shorter return-on-investment period (around 8 years) for data centers.

QHow does the article describe the shift in OpenAI's infrastructure strategy, and what is a key example?

AThe article states OpenAI has shifted from a joint venture model (like the stalled $500 billion 'Stargate' project) to a strategy of advancing its own computing power through bilateral, large-scale deals. The 20-year lease with SB Energy for the Ohio campus is cited as the largest product of this new approach.

QWhat are the significant financial commitments and associated risks for OpenAI as it prepares for its IPO?

AOpenAI faces at least $665 billion in cloud leasing commitments from deals with Microsoft, Oracle, and AWS. These long-term obligations are expected to be a major focus of SEC review and investor due diligence for its IPO, especially given internal projections of continued significant losses until at least 2029.

Пов'язані матеріали

How to Do Research Well: Deliberately Practice the Real Skills That Matter

No one truly teaches you how to do research. You're often given a desk, a pre-selected problem, and vague instructions to "create something new." Consequently, many people reverse-engineer the job based on visible outputs—papers, posts, announcements—learning only how to *appear* like a researcher rather than how to *become* one. True research capability is built from stacking small, trainable skills, nearly all of which can be developed through deliberate practice. **Pick Your Own Problem:** Most researchers absorb problems from advisors or trends, lacking the underlying reasoning. Choosing a problem you genuinely care about, as John Schulman advises, leads to original work. Develop "taste" like a muscle: predict experiment outcomes, guess paper results from methods, and track which findings remain important over time. **Upgrade Your Inputs:** Relying on shared reading lists (arXiv hot lists, filtered group chats) leads to unoriginal conclusions. Undervalued old literature often holds crucial insights (e.g., MoE, LSTM, backpropagation). Richard Sutton's "The Bitter Lesson" or Claude Shannon's 1952 talk on creative thinking are more predictive than lengthy modern surveys. Breadth matters as much as depth: draw from neuroscience, mechanism design, hardware knowledge, and honest statistics. Read papers directly, especially appendices and limitations sections. **Write Everything Down:** As Paul Graham noted, writing exposes flaws in seemingly mature ideas. Writing is the cheapest defense against self-deception. Following Feynman's principle, Darwin programmatically wrote down facts contradicting his theory to combat memory bias. Maintain a detailed log of hypotheses, setups, predictions, results, and updated understandings. Reviewing past logs fosters essential humility.

marsbit1 год тому

How to Do Research Well: Deliberately Practice the Real Skills That Matter

marsbit1 год тому

Following US Ban on Fable 5, Zhipu AI's Stock Soars 47%

On June 15th, shares of Zhipu AI surged dramatically on the Hong Kong stock market, peaking at a 47.6% gain before closing 32.82% higher. This sharp increase was directly triggered by two recent industry events. On June 12th, Anthropic announced it was suspending global access to its latest flagship models, Claude Fable 5 and Claude Mythos 5, to comply with a U.S. government export control order. The next day, Zhipu AI announced it would open access to its latest open-source flagship model, GLM-5.2, under the permissive MIT license. The Anthropic incident highlighted a critical issue beyond raw model capability: the risk of sudden, unpredictable loss of access to advanced AI models, especially for developers and enterprises deeply integrated with them. This has shifted industry and market focus toward factors like stability, sustainable access, and controllability. Zhipu's move, promoting "frontier intelligence for all," positions its openly available model as a reliable and accessible alternative. The GLM-5.2 model emphasizes "Long Horizon Task" capabilities with a 1M context window, targeting complex, multi-step coding and engineering workflows where maintaining context is crucial. Analysts note this event exposes the risk of dependency on closed-source models subject to single jurisdictional controls, potentially accelerating a shift toward domestic base models and localized deployments. The market's reaction signals a new valuation dimension in AI: providers who can offer stable, long-term, and sustainably accessible AI capabilities are gaining strategic importance.

marsbit2 год тому

Following US Ban on Fable 5, Zhipu AI's Stock Soars 47%

marsbit2 год тому

Торгівля

Спот
Ф'ючерси

Популярні статті

Як купити ADA

Ласкаво просимо до HTX.com! Ми зробили покупку Cardano (ADA) простою та зручною. Дотримуйтесь нашої покрокової інструкції, щоб розпочати свою криптовалютну подорож.Крок 1: Створіть обліковий запис на HTXВикористовуйте свою електронну пошту або номер телефону, щоб зареєструвати обліковий запис на HTX безплатно. Пройдіть безпроблемну реєстрацію й отримайте доступ до всіх функцій.ЗареєструватисьКрок 2: Перейдіть до розділу Купити крипту і виберіть спосіб оплатиКредитна/дебетова картка: використовуйте вашу картку Visa або Mastercard, щоб миттєво купити Cardano (ADA).Баланс: використовуйте кошти з балансу вашого рахунку HTX для безперешкодної торгівлі.Треті особи: ми додали популярні способи оплати, такі як Google Pay та Apple Pay, щоб підвищити зручність.P2P: Торгуйте безпосередньо з іншими користувачами на HTX.Позабіржова торгівля (OTC): ми пропонуємо індивідуальні послуги та конкурентні обмінні курси для трейдерів.Крок 3: Зберігайте свої Cardano (ADA)Після придбання Cardano (ADA) збережіть його у своєму обліковому записі на HTX. Крім того, ви можете відправити його в інше місце за допомогою блокчейн-переказу або використовувати його для торгівлі іншими криптовалютами.Крок 4: Торгівля Cardano (ADA)Легко торгуйте Cardano (ADA) на спотовому ринку HTX. Просто увійдіть до свого облікового запису, виберіть торгову пару, укладайте угоди та спостерігайте за ними в режимі реального часу. Ми пропонуємо зручний досвід як для початківців, так і для досвідчених трейдерів.

991 переглядів усьогоОпубліковано 2024.12.10Оновлено 2026.06.02

Як купити ADA

Обговорення

Ласкаво просимо до спільноти HTX. Тут ви можете бути в курсі останніх подій розвитку платформи та отримати доступ до професійної ринкової інформації. Нижче представлені думки користувачів щодо ціни ADA (ADA).

活动图片