NVIDIA and OpenAI have announced a landmark strategic partnership, with NVIDIA committing up to $100 billion in investment to support OpenAI’s next-generation AI infrastructure. The deal centers on deploying at least 10 gigawatts of NVIDIA computing systems to power OpenAI’s future models, including those on the path to superintelligence. The first phase, expected to come online in the second half of 2026, will leverage NVIDIA’s Vera Rubin platform, a next-generation architecture featuring advanced CPUs and networking. This collaboration marks a pivotal moment in the AI race, reinforcing the critical role of compute infrastructure in driving innovation. Jensen Huang, NVIDIA’s CEO, emphasized the long-standing synergy between the two companies, tracing their partnership from early DGX supercomputers to the rise of ChatGPT. “This investment and infrastructure partnership mark the next leap forward — deploying 10 gigawatts to power the next era of intelligence,” he said. Sam Altman, OpenAI’s CEO, echoed this sentiment, stating that “compute infrastructure will be the basis for the economy of the future” and that the partnership will enable broader access to AI breakthroughs. The investment will fund the construction of massive AI data centers, with the 10-gigawatt target equivalent to roughly 4 to 5 million GPUs—more than NVIDIA expects to ship globally in a single year. This underscores the immense scale of the infrastructure build-out underway across the tech industry. Microsoft, Meta, Alphabet, and Amazon are all investing heavily in AI hardware and energy capacity, with Meta’s Mark Zuckerberg warning that the greater risk lies in moving too slowly rather than overspending. Energy remains a critical challenge. As data centers grow, so does their power demand. Companies like Amazon, Google, and Microsoft are securing long-term deals with nuclear energy providers to meet these needs, though such projects have sparked local opposition and environmental concerns. Meanwhile, NVIDIA continues to dominate the AI chip market. In its latest earnings, the company revealed that two unnamed customers—widely believed to be OpenAI and another major tech firm—accounted for 39% of its revenue. This highlights the deep reliance of AI companies on NVIDIA’s hardware. In a separate but related development, NVIDIA has released a self-paced workshop on building Agentic RAG (Retrieval-Augmented Generation) agents using its Nemotron models and DevX tools. Agentic RAG enhances traditional RAG by integrating autonomous reasoning and tool use, enabling AI systems to dynamically retrieve information, make decisions, and adapt to complex tasks. The workshop guides developers through setting up a secure environment, ingesting and splitting data, embedding documents into a FAISS vector database, and creating a ReAct agent that intelligently decides when to retrieve information. It also covers migration to local NIM microservices for production use, offering full control and performance. The workshop emphasizes best practices in prompt engineering, retrieval optimization, and agent configuration, using tools like LangChain, LangGraph, and NVIDIA’s NIM inference microservices. Developers can test their agents via a Streamlit interface and trace execution for debugging. Together, these developments illustrate a transformative moment in AI: massive infrastructure investments are fueling rapid innovation, while new frameworks like Agentic RAG are making AI systems smarter, more autonomous, and better at handling real-world complexity. As the race for AI supremacy intensifies, the convergence of hardware, software, and strategic partnerships is shaping the future of intelligence.