The artificial intelligence industry is facing a new setback: Nvidia plans to contribute up to $100.000 billion to OpenAI through a strategic agreement that seeks to accelerate the deployment of data centers and the energy capacity needed to run the next batch of AI models.
Both companies have signed a letter of intent that lays the foundations for the agreement, with a view to at least 10 gigawatts (GW) of Nvidia systems integrated into OpenAI's infrastructure. The commitment will close in the coming weeks, and the schedule begins with the first phase in the second half of 2026.
An unprecedented agreement: amount, structure and schedule
The operation will be structured in stages: the initial injection of 10.000 billion dollars This will happen when the first gigawatt of computing power comes online. Nvidia will receive a stake in OpenAI, although the size of that package has not been detailed.
The deployment will include data center capacity and power supply, with the goal of train and run the next generation of models of the company headed by Sam Altman. The agreement provides for the activation of the Vera Rubin platform from Nvidia in the first wave, scheduled for the second half of 2026.
The parties emphasize that this is a preferential alliance in computing and networks, with joint work to optimize architecture, scaling and performance of the AI ​​infrastructure.
In parallel, the document clarifies that the final terms will be finalized shortly, maintaining the implementation roadmap in stages to keep pace disbursements, deliveries and capacity targets.
Large-scale infrastructure: 10 GW and Nvidia's next platform
The agreement includes the deployment of at least 10 GW of Nvidia systems for OpenAI, a magnitude that, according to industry estimates, is equivalent to between 4 and 5 million GPUs dedicated to AI. The technical ambition is to build a network of next-generation data centers with fine-tuned efficiency, interconnection, and power supply.
The first phase will operate on Vera Rubin, the ecosystem that will succeed Nvidia's current platform and is designed for large-scale training and inference workloads. The stated goal is to pave the way for AI systems. increasingly powerful and versatile.
In line with this roadmap, Nvidia and OpenAI plan to integrate their work with a network of technology partners, including Microsoft, Oracle, SoftBank and collaborators of the Stargate program, to assemble a high-performance infrastructure.
Although the core of the agreement is technical, it also has clear financial aspects: Nvidia gains equity exposure at OpenAI and, at the same time, a backlog of orders tied to the deployment of its own systems, while OpenAI ensures supply and financing aligned with its product schedule.
Market impact and strategic keys
The stock market reaction was immediate: Nvidia shares rose by around 4% intraday, approaching historic highs and consolidating its position among the world's most valuable companies by market capitalization.
From a competitive perspective, the movement seeks strengthen the role of Nvidia hardware as the backbone of reference AI systems, at a time when large technology companies are exploring their own chip designs. OpenAI, for example, works with Broadcom on alternatives for internal use.
Sam Altman and Jensen Huang agree that the computing power will support the economy in the coming years: hence the emphasis on scaling energy, networks and data centers to accompany the training of increasingly complex models, including the incorporation of Google Cloud.
OpenAI claims to have surpassed 700 million active users per month, a reference that helps understand the urgency of expanding infrastructure to support current services and open the door to new capabilities.
Overall, the operation aligns incentives: milestone financing, guaranteed supply, adoption of next-generation Nvidia systems, and coordinated scaling priorities with key ecosystem partners.
The announcement sets out a demanding roadmap: closing of the final agreement in weeks, activation of the first phase in the second half of 2026 and a 10 GW deployment that, due to its volume, will test the supply chain and the integration capacity of all those involved.
The commitment to massive infrastructure, cross-shareholdings, and a timeline structured by stages sends a clear message: Nvidia and OpenAI accelerate their collaboration to turn the performance leap into a sustained advantage by combining capital, technology, and execution.
