Plans to expand a major artificial intelligence data center in Texas have been abandoned after negotiations between Oracle and OpenAI stalled, highlighting the enormous financial complexity behind building next-generation computing infrastructure.
The proposed expansion had been linked to the ambitious Stargate initiative, a massive project designed to develop up to ten gigawatts of AI computing capacity through an estimated $500 billion investment program.
Originally unveiled in January 2025 with support from SoftBank and other technology partners, the initiative was intended to accelerate the construction of large-scale data centers capable of powering advanced artificial intelligence services.
However, discussions surrounding the Texas expansion broke down as financing arrangements became increasingly complicated and OpenAI’s infrastructure requirements evolved during the lengthy negotiation process.
Industry sources indicated that the abandoned expansion centered on an additional 600 megawatts of potential computing capacity that had been proposed near the primary Stargate campus in Abilene.
Existing Abilene Facilities Continue Operating
Despite the cancellation of the expansion plan, construction activity at the main Stargate site in Abilene continues, with several facilities already operational and additional capacity expected to come online gradually.
The campus currently includes eight buildings intended to house large-scale computing clusters, two of which are already running and supporting artificial intelligence workloads managed by Oracle Cloud Infrastructure.
Technology companies across the industry have dramatically increased spending on such facilities because generative artificial intelligence models require immense processing power and specialized hardware to operate effectively.
Platforms such as ChatGPT and Microsoft’s Copilot rely on massive server farms filled with advanced graphics processing units capable of handling the enormous computational requirements associated with training and running modern AI systems.
Even with the expansion canceled, Oracle and OpenAI are reportedly continuing plans to develop an additional 4.5 gigawatts of data center capacity elsewhere within the broader Stargate network.
Meta Explores Opportunity At Abilene Site
The collapse of negotiations between Oracle and OpenAI appears to have opened the door for other technology companies seeking access to large-scale computing infrastructure in the rapidly expanding artificial intelligence market.
Reports indicate that Meta Platforms has begun exploring the possibility of leasing the previously planned expansion site in Abilene from infrastructure developer Crusoe, which has been involved in building data centers across the region.
The discussions were reportedly facilitated by Nvidia, whose advanced artificial intelligence processors form the backbone of many modern data center deployments across the technology sector.
According to industry sources, Nvidia played a role in encouraging discussions to ensure its own semiconductors would power the expanded facility rather than chips produced by rival manufacturer Advanced Micro Devices.
Meta declined to comment publicly on the potential leasing arrangement, while OpenAI and Nvidia did not immediately respond to inquiries regarding the negotiations.
AI Infrastructure Race Accelerates Globally
The evolving plans around Stargate highlight the increasingly competitive race among technology giants to secure the infrastructure necessary to support artificial intelligence services used by hundreds of millions of people worldwide.
Building AI-ready data centers requires enormous capital investment, vast energy supplies, and advanced cooling systems capable of managing the extreme heat generated by densely packed computing hardware.
As companies accelerate spending to gain an advantage in artificial intelligence development, large-scale projects like Stargate are becoming critical battlegrounds for technological leadership and future digital infrastructure.
