#OpenAI builds 5 computing centers at once! Nvidia feeds Masayoshi Son and oracle bone inscriptions
Nvidia has just planned to invest $10 million in OpenAI, and OpenAI has announced the use of the money: it will collaborate with Oracle and SoftBank to build data centers, with five in one go.
This batch of newly built data centers will be part of Ultraman’s “Stargate” program, and the planned capacity of the project will be increased to nearly 7GW, roughly equivalent to seven large nuclear reactors.
OpenAI stated that the new data center will advance the completion of the Stargate project.
OpenAI purchases Oracle’s cloud services → Oracle buys Nvidia GPU → Nvidia invests in OpenAI. With the announcement of the new plan, this triangular relationship has been further strengthened.
In addition, Ultraman also revealed on his personal blog that OpenAI will have new plans and partners, as well as some “new ideas” in terms of computing power construction in the coming months.
But I don’t know if you’ve noticed in this collaboration, but Microsoft, the former big brother, is no longer a big deal.
Microsoft falls, Oracle eats.
##Collaborate with Oracle to build a data center
The five new data centers announced by OpenAI will be developed in collaboration with Oracle and SoftBank.
Three of them were built in collaboration with Oracle, located in Shackelford County, Texas, Danana County, New Mexico, and an undisclosed location in the Midwest.
The other two are operated by SB Energy, a subsidiary of OpenAI and SoftBank that develops solar and battery projects, located in Loztown, Ohio and Milam County, Texas, respectively.
Of course, the collaboration between OpenAI and Oracle is already well-established. The data center in Abilene, Texas was jointly built by the two companies, and an agreement was reached in July to expand the data center and develop up to 4.5GW of data center capacity.
The operating model is “lease for sale” - the data center located in Abilene is mainly owned and operated by Oracle, while OpenAI is used as the main tenant.
According to sources, the data center is currently running on Oracle Cloud infrastructure and providing support for OpenAI’s training and inference workloads.
Oracle is currently building eight data center halls in Abilene, each with a capacity of approximately 100MW. After completion, the facility will accommodate over 400000 GPUs, reaching a scale of 1.4GW.
##NVIDIA invests billions of dollars in OpenAI
As another aspect of the ‘triangle relationship’, Nvidia just announced yesterday that it plans to invest $100 billion in OpenAI to build a 10GW data center, which is roughly equivalent to 4-5 million GPUs.
However, this $100 billion is not a one-time investment, but will be made in batches. OpenAI invests $10 billion for every 1GW facility built, with the first batch planned to be completed in the second half of next year.
100 billion US dollars, one-fifth of OpenAI’s valuation, is indeed a considerable amount, but is it really enough for OpenAI?
During the August earnings conference call, Huang Renxun stated that the cost of building an artificial intelligence data center with a capacity of approximately 50 billion US dollars per GW.
But Nvidia’s average investment in GW is only 10 billion, and it is only given after completion. Where should the remaining money come from has become a major question for people.
OpenAI did not respond to requests for comment regarding its financing plan, and neither Nvidia nor OpenAI have provided a specific timeline for the launch of all 10 GW of production capacity.
Prior to Nvidia’s announcement of investment, debt rating agency Moody’s had also raised similar concerns, questioning whether OpenAI had sufficient cash flow to pay for Oracle’s contract.
However, Nvidia’s investment news injected a shot in the arm for OpenAI, increasing the certainty that OpenAI will purchase cloud resources from Oracle as planned.
One More Thing
OpenAI initially positioned Stargate as a “new company” and will be chaired by SoftBank CEO Masayoshi Son.
However, according to executives close to the project, Stargate is actually a general term used to refer to all OpenAI data center projects.
However, the executive specifically pointed out that this does not include projects developed in collaboration with Microsoft.
As a long-standing investor in OpenAI, Microsoft seems to have been excluded from the blueprint of the Stargate project.
Microsoft and OpenAI, the honeymoon period is over.
Reference link:
[1] https://openai.com/index/openai-nvidia-systems-partnership/
[2] https://www.wired.com/story/openai-oracle-softbank-data-center-stargate-us/
[3] https://blog.samaltman.com/abundant-intelligence
[4] https://www.reuters.com/business/more-questions-than-answers-nvidias-100-billion-openai-deal-2025-09-23/