NVIDIA just made its biggest infrastructure play yet. On August 17, the company announced a partnership with SB Energy to secure land, power, and shell (LPS) capacity at the PORTS-Pike Technology Campus in Portsmouth, Ohio. OpenAI will be the tenant. The initial deployment covers 4.25 gigawatts of AI factory capacity, with an option to expand to 8 gigawatts. OpenAI's total commitments now represent roughly 12 gigawatts of NVIDIA compute through 2030, with room to reach 16 gigawatts. That's approximately $600 billion in NVIDIA compute revenue over the next four years.
The move marks a shift in how NVIDIA operates. The company is no longer just selling chips. It's using its balance sheet to lock in the physical resources that AI data center operations need: land, electricity, and buildings. LPS has become the critical constraint on AI factory deployment, and NVIDIA is treating it the same way it treats semiconductor supply chain management.
Key Updates at a Glance
Here's a breakdown of the NVIDIA AI infrastructure announcements from August 2026, covering partnership details, capacity targets, and financial commitments:
| Category | Update |
|---|---|
| Partnership | NVIDIA + SB Energy + OpenAI at PORTS-Pike, Ohio |
| Initial Capacity | 4.25 GW (expandable to 8 GW) |
| NVIDIA Investment | $1.5 billion equity stake in SB Energy |
| OpenAI Commitment | ~12 GW of NVIDIA compute through 2030 |
| Total Value | ~$600 billion in NVIDIA compute opportunity |
| Timeline | Phased online 2028 to 2030 |
| Site Lease | 20-year term, OpenAI as tenant |
| Financing Platform | $500B+ with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, KKR |
| Lancium Investment | Up to $3 billion (20-30% equity stake) |
NVIDIA's PORTS-Pike Partnership: Securing LPS for AI Factories
The core announcement is straightforward. SB Energy, backed by SoftBank, will develop and operate the PORTS-Pike campus under a 20-year lease with OpenAI. NVIDIA provides credit support for portions of lease and power payments, plus a residual-value commitment. NVIDIA also invested $1.5 billion in SB Energy directly.
The site matters. PORTS-Pike is a former federal facility that once housed a gaseous diffusion plant in Pike County, Ohio. SB Energy and SoftBank plan to build at least 10 gigawatts of new power generation capacity, with $4.2 billion in regional grid upgrades. The first 800 megawatts could come online as early as 2028 using existing grid infrastructure. Not a small number. But just the start.
NVIDIA's blog post, "Securing the Infrastructure of Intelligence," frames LPS as the next strategic resource. For most NVIDIA customers, securing LPS is already part of their infrastructure strategy. Large cloud providers have the balance sheets and credit profiles to handle it independently. But frontier AI labs are different. They have extraordinary demand for training and inference compute but grow faster than their balance sheets can support. Their growth is constrained not by algorithms or customer demand, but by the availability of compute. NVIDIA bridges that gap by securing the physical site and letting the AI lab focus on what it does best.
The guarantee structure is phased. It becomes effective as data centers come online between 2028 and 2030. As OpenAI makes lease payments and capacity ramps up, NVIDIA's remaining exposure declines. NVIDIA's support is limited to defined portions of lease and power payments, not the full cost of the site or all of the tenant's obligations.
OpenAI's $600 Billion Compute Commitment Through 2030
The numbers are big. OpenAI has committed to substantial deployments of NVIDIA AI infrastructure through 2030. Existing and planned commitments represent approximately 12 gigawatts of NVIDIA compute. If NVIDIA extends the PORTS-Pike arrangement beyond the initial 4.25 gigawatts, that figure could reach 16 gigawatts. At those levels, the OpenAI infrastructure investment represents roughly $600 billion of NVIDIA compute through 2030.
Each generation of NVIDIA AI factory systems deployed at PORTS-Pike could represent approximately 1.5 million NVIDIA GPUs, or $150 billion to $200 billion in NVIDIA revenue. Over 20 years, the site can support multiple upgrade cycles. The LPS commitment secures a long-lived AI factory site. The compute inside can be upgraded repeatedly. Each new generation delivers greater production, more intelligence, better economics.
This matters because the AI compute capacity crunch is real. NVIDIA's Blackwell GPUs are in such high demand that supply remains constrained. The Vera Rubin platform is now in full production, with Taiwan's top server makers shipping systems at scale. But getting enough power and land to deploy these systems is the harder problem. OpenAI's Stargate initiative, originally announced as a $500 billion plan in January 2025, has faced construction delays, financing disputes, and local opposition at multiple sites. The PORTS-Pike deal gives OpenAI a secured path to capacity without the financing friction that has slowed other Stargate projects.
Image source: Nvidia
NVIDIA's $500B AI Data Center Financing Platform
The PORTS-Pike deal isn't happening in isolation. On August 8, NVIDIA announced strategic partnerships with six major financial institutions to establish independent compute financing platforms. The partners are Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR. The goal: mobilize over $500 billion of third-party capital to support AI data center buildout. NVIDIA AI Factory Compute is being positioned as an investable asset class.
That same week, NVIDIA agreed to invest up to $3 billion in Lancium, a power infrastructure developer backed by Blackstone. The first $2 billion buys roughly 20% equity. An additional $1 billion is contingent on Lancium securing more power connections. Lancium is the company behind the Stargate data center campus in Abilene, Texas, where OpenAI and Oracle already operate facilities. The investment is pure equity, no credit guarantees attached.
NVIDIA's 2026 investment portfolio tells the broader story: $30 billion into OpenAI, $2 billion each into CoreWeave and Nebius, $1.5 billion into SB Energy, up to $3 billion into Lancium, $3.2 billion into Corning for optical fiber, and $2.1 billion into IREN for data center operations, with total private company equity holdings reaching $22.25 billion as of the latest fiscal year-end, up from $3.39 billion a year earlier.
Semiconductor Supply Chain Pressure: Blackwell Demand and Beyond
The semiconductor supply chain is feeling the strain. NVIDIA's Blackwell GPUs are sold out. Analysts at Wedbush report that customers face longer delivery times and harder acquisition paths. BNP Paribas Exane notes that 2026 demand far outpaces supply, with NVIDIA expected to ship $320 to $340 billion in Blackwell and Rubin products this calendar year.
The bottleneck has shifted. Getting chips is hard. Getting power for those chips is harder. Morgan Stanley reports that Blackwell production is actually re-accelerating, with the platform's end-of-life timeline pushed back several months because demand keeps growing. But the AI energy grid can't expand fast enough to keep up.
That's why NVIDIA is investing upstream. The Lancium deal locks in power connections in Texas. The SB Energy partnership secures a site in Ohio with dedicated power generation. For the broader supply chain, this means NVIDIA is pulling demand forward in a way that gives suppliers multi-year visibility. Corning is expanding U.S. optical fiber production. Memory suppliers like Micron are hitting record revenue on HBM4 demand for NVIDIA's Vera Rubin platform. The GPU supply chain, once measured in chip yields and wafer starts, now extends to power plants and transmission lines.
NVIDIA Infrastructure Risks: Circular Financing and Concentration
The deal raises fair questions. Is NVIDIA financing its own customers' demand? NVIDIA is using its balance sheet to support customer infrastructure, which some analysts describe as circular financing. NVIDIA's blog post addresses this directly: "No. OpenAI will pay the lease." The company frames its role as identical to its supply-chain discipline: secure critical inputs when there's visibility into customer demand and when doing so enables long-term productive capacity.
There's also the question of what happens if OpenAI doesn't use the site. NVIDIA says its compute is versatile and fungible. The capacity can be resold to another qualified tenant across NVIDIA's global network of cloud providers, enterprises, AI labs, and startups. CUDA makes NVIDIA compute broadly adoptable, which means the site isn't tied to one customer.
But the scale of NVIDIA's financial commitments is unprecedented. The $105 billion credit guarantee for PORTS-Pike. The $500 billion financing platform. The $30 billion OpenAI investment. The Lancium and SB Energy stakes. If AI demand falls short of projections, NVIDIA's exposure is real. The company acknowledges this by structuring guarantees to phase out as capacity comes online and OpenAI makes lease payments.
NVIDIA has moved from building chips to securing the full stack of AI infrastructure. The PORTS-Pike partnership with SB Energy and OpenAI locks in 8 gigawatts of AI compute capacity in Ohio, part of a 12 to 16 gigawatt commitment through 2030 that could represent $600 billion in NVIDIA compute revenue. Combined with $500 billion in financing partnerships and multi-billion-dollar stakes in power infrastructure companies, NVIDIA is betting its own capital that AI compute demand will keep scaling for decades. The first 800 megawatts at PORTS-Pike could be operational by 2028. By then, we'll know if the bet is paying off.
* Learn more on the Nvidia official site: https://nvidianews.nvidia.com/news/openai-and-nvidia-announce-strategic-partnership-to-deploy-10gw-of-nvidia-systems
- Jak pobrać najnowszą wersję Pokémon GO APK 0.319.0 na Androida 2024
- Jak pobrać najnowszą wersję Chapters APK 6.5.7 na Androida 2024
- Jak pobrać PK XD: Fun, friends & games APK Najnowsza wersja 1.50.4 dla Androida 2024