According to a press release issued by NVIDIA, OpenAI will build and operate a world-class artificial intelligence factory at PORTS-Pike. The AI factory will utilize NVIDIA’s full-stack DSX AI factory platform, including GPUs, CPUs, networking, and infrastructure software. The initial deployment is expected to provide 4.25 gigawatts of AI factory capacity. Each generation of the NVIDIA AI factory system deployed at PORTS-Pike represents approximately 1.5 million NVIDIA GPUs, corresponding to roughly USD 150 billion to USD 200 billion in NVIDIA revenue. After 20 years of development, the site supports multiple upgrade cycles. NVIDIA may also choose to extend the PORTS-Pike arrangement beyond the initial 4.25 gigawatts to secure the remaining 3.75 gigawatts of capacity.
OpenAI has committed to the large-scale deployment of NVIDIA AI infrastructure by 2030. OpenAI’s existing and planned commitments account for approximately 12 gigawatts of NVIDIA computing capacity, with the potential to expand to around 16 gigawatts if NVIDIA extends the PORTS-Pike agreement accordingly. NVIDIA stated that, at these levels, this opportunity represents approximately USD 600 billion in NVIDIA computing revenue by 2030.
AI factories are decisive infrastructure in the AI era—where computing power transforms energy and data into intelligence, driving business operations across industries and nations.
In the AI economy, computing power equates to revenue.
AI factories require a comprehensive suite of critical resources: advanced chips, packaging components, memory, networking equipment, as well as land, power, and shell building support (LPS).
Just as$NVIDIA (NVDA.US)$The company is now applying the same strategy to secure Land, Power, and Shell (LPS) resources for NVIDIA’s AI factories, leveraging its scale, long-term business forecasting, and supply chain partnerships—much like it did to acquire critical semiconductor resources.
Today, the company has partnered with SB Energy to secure LPS resources at the highly advantageous PORTS-Pike Technology Park in Portsmouth, Ohio, for deploying NVIDIA computing power, with OpenAI serving as the tenant for the site.
As of the time of writing,$NVIDIA (NVDA.US)$It rose slightly by 0.17% during trading hours to close at $225.545, with a turnover of $1.369 billion.
LPS: The Next Class of Strategic Resources
For the vast majority of$NVIDIA (NVDA.US)$customers, securing land, power, and shell infrastructure (LPS) has long been an integral part of their infrastructure strategy.
Global large-scale cloud service providers and investment-grade enterprises possess robust balance sheets and infrastructure expertise, enabling them to independently secure LPS resources through long-term contracts. These companies build and operate AI factories leveraging NVIDIA’s accelerated computing, networking, systems, and software products.
This business model will continue to account for the majority of NVIDIA’s business.
However, the situation differs for frontier AI laboratories.
Frontier AI laboratories have substantial demands for computational power for both training and inference. However, the growth pace of many institutions has already exceeded what their balance sheet size and long-term creditworthiness can support. While these entities may enjoy robust customer demand and rapid revenue growth, they still lack the decades-long infrastructure contracts and investment-grade financing capabilities necessary to independently build AI factory infrastructure.
The constraint on their development is often no longer algorithms or customer demand, but rather the supply of computational power.
For such enterprises, greater computational capacity translates into enhanced intelligence capabilities, a richer product portfolio, a larger user base, and higher revenues. NVIDIA is helping to provide the infrastructure that drives this virtuous cycle.
PORTS-Pike: A site capable of hosting multiple generations of NVIDIA computational power
OpenAI will build and operate a world-class AI factory at PORTS-Pike. This AI factory will adopt NVIDIA’s full suite of DSX AI Factory platforms, including GPUs, CPUs, networking equipment, and infrastructure software.
The initial deployment is expected to provide 4.25 gigawatts of AI factory computational power. Each generation of NVIDIA AI Factory systems deployed at PORTS-Pike corresponds to approximately 1.5 million NVIDIA GPUs, representing an estimated NVIDIA revenue of USD 150 billion to USD 200 billion. The site can support multiple rounds of hardware upgrades and iterations over a 20-year cycle.
This has core economic significance: investments related to Long-Term Power Supply (LPS) secure long-term usable AI factory sites, while the NVIDIA computational equipment within these sites can be iteratively upgraded. Each new generation of systems delivers higher output, stronger intelligence capabilities, and improved economic efficiency.
NVIDIA also has the option to expand the scope of its partnership at PORTS-Pike to secure the remaining 3.75 gigawatts of computational power resources, in addition to the initial 4.25 gigawatts.
Opportunities for OpenAI and NVIDIA to Expand Computational Power Deployment
Looking at longer-term plans, OpenAI has committed to large-scale deployment of NVIDIA AI infrastructure by 2030. OpenAI’s current and planned computational power requirements total approximately 12 gigawatts of NVIDIA capacity; if NVIDIA expands the PORTS-Pike partnership beyond the initial 4.25 gigawatts, the total computational scale could reach up to approximately 16 gigawatts.
Based on this scale, the partnership is projected to generate approximately $600 billion in computing-related business opportunities for NVIDIA by 2030.
Key Questions and Answers
$NVIDIA (NVDA.US)$What does the guarantee cover, and what is its validity period?
NVIDIA will provide 20 years of support for the LPS infrastructure at PORTS-Pike, with a capacity of approximately 4 gigawatts, securing a site dedicated to the deployment of NVIDIA computing power.
The company's support is limited to certain agreed-upon rental payments and electricity costs, as well as a specific residual value commitment. It does not cover the full cost of the site nor assume all obligations of the lessee.
The guarantee will take effect gradually as the data centers become operational in phases between 2028 and 2030. As OpenAI pays rent and computing capacity comes online, NVIDIA's potential risk exposure will decrease accordingly.
$NVIDIA (NVDA.US)$Why provide a guarantee for PORTS-Pike?
Supporting infrastructure for large-scale power supply (LPS) has become a key bottleneck constraining the establishment of AI factories. NVIDIA will selectively secure high-quality sites capable of hosting multiple generations of NVIDIA computing systems and meeting sustained, stable customer demand.
The effective service life of the site can span multiple generations of NVIDIA systems, with each generation delivering stronger intelligence capabilities and generating higher revenue than its predecessor.
Does this constitute circular financing?
That is not the case. The rent will be borne by OpenAI.
NVIDIA secured PORTS-Pike, leveraging its business scale and long-term demand forecasts, to deploy NVIDIA computing power. This aligns with the company’s supply chain management logic: when customer demand can be clearly anticipated and related investments yield long-term available capacity, key production factors are locked in.
If OpenAI ceases to use the PORTS-Pike site in the future, how will the project be handled?
$NVIDIA (NVDA.US)$Computing power is generic and transferable, and has been widely adopted across the industry. These computing resources can be resold to other qualified lessees within NVIDIA’s global ecosystem, including cloud service providers, enterprises, AI laboratories, and startups.
CUDA makes NVIDIA computing power more than just hardware. CUDA provides a unified platform for developers and NVIDIA engineers, enabling continuous iterative optimization of deployed systems.
CUDA endows NVIDIA computing power with general-purpose capabilities; this generality enables transferability; and transferability further enhances resource utilization and asset retention value, making NVIDIA computing power an asset with tangible productive capacity that can be leased out or used for financing.
The value of high-quality sites like PORTS-Pike is not limited to a single customer or a single generation of computing hardware. NVIDIA’s standardized platform, vast developer ecosystem, and large base of potential users ensure that the computing capacity at such sites can be reallocated for future use.
$NVIDIA (NVDA.US)$What is the planned scale of LPS resources to be acquired?
The company will proceed prudently with these efforts, guided by strategic considerations.
The vast majority of NVIDIA customers will continue to procure the LPS resources they need independently. Most LPS infrastructure supporting NVIDIA computing power will still be directly implemented by cloud service providers, enterprises, national-level AI builders, and other customers.
NVIDIA will selectively secure high-quality sites that demonstrate clear, sustainable demand capable of supporting multiple generations of NVIDIA computing deployments.
Infrastructure for the Intelligent Era
PORTS-Pike represents$NVIDIA (NVDA.US)$Another step forward in the development process.
NVIDIA initially developed accelerated computing chips, later expanding its business to include systems, networking, CUDA, and full-stack AI factories. Today, the company is actively involved in securing the critical infrastructure necessary to build such facilities.
NVIDIA is a full-stack AI infrastructure platform.
The company establishes a robust, long-term foundation for AI factories, enabling customers to deploy globally leading computing platforms across successive generations.
By securing key resources required for deploying NVIDIA computing capabilities, NVIDIA empowers the world’s most innovative enterprises to build AI factories, driving progress in the Intelligent Era.
Meanwhile,$NVIDIA (NVDA.US)$It was also announced that a $1.5 billion investment would be made in SB Energy to support its growth and commitment to Ohio. NVIDIA announced that, through its partnership with SB Energy, it has secured land, power, and shell (LPS) capacity at the PORTS-Pike Technology Park in Pike County, Ohio, to host NVIDIA computing workloads. OpenAI will be the customer. SB Energy will build, own, and operate the data center under a 20-year lease agreement with OpenAI. The PORTS-Pike campus project is expected to create tens of thousands of jobs in Ohio, fund electricity infrastructure construction, and invest hundreds of millions of dollars in the community, anchored by an initial $80 million community benefit fund. NVIDIA will provide credit support for the construction of land, power, and shell facilities to secure an initial 4.25 IT-GW, with an option to acquire the remaining 3.75 IT-GW.
Source: NVIDIA Official Blog