llms.blog
AI

OpenAI signs an 8-gigawatt Ohio data center lease backed by up to $105B from Nvidia

OpenAI has signed a 20-year lease for the PORTS-Pike data center campus in Ohio, its largest infrastructure commitment to date, with Nvidia agreeing to backstop up to $105 billion of the project's value. The deal, struck with SoftBank's SB Energy, gives OpenAI roughly 8 gigawatts of IT capacity, about 10 gigawatts once cooling and supporting infrastructure are included. The site sits partly on a former U.S. Department of Energy uranium enrichment facility and draws power from a 9.2-gigawatt gas

axc
axc
2 min
OpenAI signs an 8-gigawatt Ohio data center lease backed by up to $105B from Nvidia

OpenAI has signed a 20-year lease for the PORTS-Pike data center campus in Ohio, its largest infrastructure commitment to date, with Nvidia agreeing to backstop up to $105 billion of the project's value.

The deal, struck with SoftBank's SB Energy, gives OpenAI roughly 8 gigawatts of IT capacity, about 10 gigawatts once cooling and supporting infrastructure are included. The site sits partly on a former U.S. Department of Energy uranium enrichment facility and draws power from a 9.2-gigawatt gas plant owned by the U.S. government and financed by Japan as part of a trade agreement.

Abstract diagram of a power plant feeding an Ohio data center full of GPU racks

Nvidia becomes the backstop

Nvidia is not guaranteeing OpenAI's rent. Instead, it is backing the residual value of the finished data centers in the first construction phase, which covers 4.25 gigawatts of IT capacity. If OpenAI walks away, SB Energy must first find a replacement tenant and attempt a sale before Nvidia covers any shortfall, capped at $105 billion. In return, Nvidia becomes the exclusive chip supplier for the first half of the site and is investing $1.5 billion in SB Energy.

Nvidia CEO Jensen Huang said the real bottleneck in the AI buildout has shifted from chips and networking to what he called LPS, meaning land, power, and shell: the site, the electricity supply, and the building itself. He expects about 1.5 million GPUs per system generation from the site, worth an estimated $150 billion to $200 billion in revenue.

Off the balance sheet

A Wall Street Journal analysis cited in the reporting shows how large these commitments have become. Nine technology companies, including Alphabet, Meta, Microsoft, and Nvidia, hold around $3 trillion in mostly AI-related obligations that do not appear on their balance sheets. Leases are recorded only once payments begin, and purchase commitments only once goods are delivered. Leases that have not yet started now total about $1.2 trillion, four times the amount a year earlier. At Alphabet, purchase commitments rose from $332 billion to $811 billion in three months.

The contracts are nearly impossible to cancel, and both Alphabet and Amazon recently reported negative free cash flow. Morgan Stanley analysts warn that investors can barely gauge these companies' true debt levels. OpenAI says it pays only for finished capacity, with the first 800 megawatts scheduled to come online in 2028.

Sources

OpenAI signs record Ohio data center lease with Nvidia backing up to $105 billion - The Decoder, Aug 17, 2026: https://the-decoder.com/openai-signs-record-ohio-data-center-lease-with-nvidia-backing-up-to-105-billion/

OpenAI locks in lease for huge data center in Ohio with backing from Nvidia - The Wall Street Journal, Aug 17, 2026: https://www.wsj.com/articles/openai-locks-in-lease-for-huge-data-center-in-ohio-with-backing-from-nvidia-7474bb9c

OpenAI joins the PORTS-Pike project - OpenAI, Aug 17, 2026: https://openai.com/index/openai-joins-ports-pike-project/

Nvidia guarantees SB Energy's PORTS-Pike technology campus in Ohio to exclusively host Nvidia AI compute - Nvidia Newsroom, Aug 17, 2026: https://nvidianews.nvidia.com/news/nvidia-guarantees-sb-energy-s-ports-pike-technology-campus-in-ohio-to-exclusively-host-nvidia-ai-compute

AIData CentersNvidiaOpenAI
ShareX (Twitter)LinkedIn

Written by

axc

More to read

  • Grokking in Large Language Models: How Weight Decay and Circuit Efficiency Drive Delayed Generalization

    Grokking in Large Language Models: How Weight Decay and Circuit Efficiency Drive Delayed Generalization In standard machine learning paradigms, model generalization closely tracks training loss: as an optimizer minimizes loss on training data, performance on held-out validation data improves in tandem until the model begins to overfit. In 2022, researchers at OpenAI observed a phenomenon that inverted this assumption: small neural networks trained on algorithmic tasks achieved near-zero trainin

    1 min

  • Serverless GPU Inference in Production: Cold Starts, GPU Memory Snapshotting, and Weight Paging Architectures

    Serverless GPU Inference in Production: Cold Starts, GPU Memory Snapshotting, and Weight Paging Architectures Deploying large language models on dedicated cloud GPUs creates an uncomfortable financial trade-off: keeping enterprise accelerators such as NVIDIA H100s or A100s warm 24/7 costs thousands of dollars per instance each month, yet scaling instances to zero introduces severe latency penalties. When traffic arrives at a dormant node, a standard inference server cold start can take anywhere

    1 min

  • Transformer Feed-Forward Networks as Key-Value Memories: How First-Layer Keys and Second-Layer Values Store Knowledge

    Transformer Feed-Forward Networks as Key-Value Memories: How First-Layer Keys and Second-Layer Values Store Knowledge In modern autoregressive Transformers, the division of labor between attention heads and multi-layer perceptron (MLP) blocks is often summarized through a clean functional split: attention routes information across sequence positions, while feed-forward networks (FFNs) process information per position. Yet for years, the exact mechanism by which FFNs process that information rem

    1 min