Dinner Strategies: Tech Titans Ask Nvidia for GPUs

To secure the coveted Nvidia GPUs for Artificial Intelligence (AI) and machine learning, a simple dinner might be the key. Last week, Larry Ellison, co-founder and technical director of Oracle, announced that he, along with Elon Musk, arranged a dinner with Nvidia CEO Jensen Huang at Nobu Palo Alto to convince him of the urgent need for additional GPUs. "We pretty much begged him to give us more GPUs," Ellison recalled of the dinner. The exchange was successful, Ellison summarized: "It worked." Ellison, whose wealth is estimated at $206 billion, is regarded as a visionary figure in the technology sector. Through strategic decisions, he has profitably aligned Oracle with cloud-based infrastructure. Nvidia, which controls over 80 percent of the AI chip market, is an indispensable partner for Oracle and other companies engaged in the AI sector. The recent analyst meeting made it clear that the competitive landscape will not derail Oracle. Ellison compared the competition in the AI sector to a Formula 1 race, stating that ultimately, there will be only one winner. For Oracle, investment in GPU technology is essential. In the first quarter of fiscal year 2025, the company reported a revenue increase of 7 percent to $13.3 billion and a profit of $2.9 billion. Oracle plans to operate 162 cloud data centers worldwide that rely on Nvidia GPU clusters to train large-scale AI models. In the first quarter, 42 new cloud GPU contracts worth $3 billion were signed. Tesla, led by Musk, also relies on Nvidia GPUs to operate supercomputers for autonomous driving. Ellison emphasized that the development of the most advanced neural networks is of great importance. He noted that billions in investments are necessary to remain competitive. "It costs $100 billion to develop a top model over the next three years," he said. "Not many companies or countries can afford that.