June 4, 2025
Oracle -founder Larry Ellison has just delivered fantastic news for investors in Nvidia

Oracle -founder Larry Ellison has just delivered fantastic news for investors in Nvidia

Larry Ellison owns 42% of Oracle (NYSE: ORCL)A technology giant of $ 465 billion that builds some of the most powerful data centers for the development of artificial intelligence (AI).

Nvidia (Nasdaq: NVDA) Oracle and most other technology companies with Datacenter chips called Graphics Processing Units (GPUs). Nvidia has had a striking increase in income in the past year and GPU’s demand continues to exceed the supply. However, some investors started to wonder how long Oracle and his colleagues can throw billions of dollars at the chip giant to feed their AI ambitions.

Ensures that the AI ​​train may start to lose steam, an important reason is that Nvidia shares trades 14.5% compared to the all time. But this month the market may have missed comments from Ellison during the financial analyst meeting of Oracle that suggest more fantastic news for the investors of Nvidia.

Oracle data centers are unique because they are automated. Each is operational identical, regardless of its size, and because they do not require human employees, it enables the company to build them quickly. Plus, Oracle’s RDMA (Random Direct Memory Access) GPU network technology can flow more quickly from one point to another than traditional Ethernet networks.

Since most AI developers pay per minute for computer capacity, the data centers of Oracle can result in considerable cost savings compared to competing infrastructure. That is why the demand rises from leading AI startups such as OpenAi, Cohere and Xai. Oracle had 85 data centers in motion with 77 more under construction from the first quarter of 2025 (ending on August 31), but Ellison thinks it could work no less than 2,000 in the long term.

Next year Oracle is planning to offer a cluster of 131,072 GPUs, which is a big step than the largest clusters now, around 32,000 GPUs. But there is another difference: the new cluster will use the newest Blackwell -Chips from Nvidia, which can perform AI insertion with 30 times the pace of its flagship H100, which Oracle currently uses. Theoretically, developers will enable developers to build the largest AI models in history.

Nvidia will consider that considerably. It generated $ 26.3 billion in data center income during his tax 2025 second quarter (ending on July 28) mainly from the GPU turnover, which was an increase of 154% compared to the period of the year ago. That growth rate delayed compared to previous quarters because the figures have become so large, but Nvidia’s customers do not display signs of withdrawal.

Leave a Reply

Your email address will not be published. Required fields are marked *