As the demand for artificial intelligence continues to skyrocket, OpenAI, the company behind the immensely popular ChatGPT, finds itself in need of additional computing power to keep up with its rapidly growing user base. Despite Microsoft's ongoing efforts to dedicate additional infrastructure to support OpenAI's needs, it has become evident that even the tech giant's resources alone are not sufficient to meet the demand.
To be clear, this isn't a problem unique to OpenAI. As AI models become larger, more complex, and data-intensive, the need for robust infrastructure to train and run these models has never been greater. This has prompted major hyperscalers like Oracle, Microsoft, and Google to forge strategic partnerships and make unprecedented multicloud arrangements.
At their inaugural Dev Day last November, OpenAI CEO, Sam Altman, shared that ChatGPT had 100 million weekly active users. This number is expected to skyrocket following Apple's announcement on Monday that it will integrate ChatGPT into all its platforms. For context, Apple’s active device base surpassed 2.2 billion in February. This will significantly expand the reach of OpenAI’s services and undoubtedly put strain on Microsoft's infrastructure.
Keep in mind that OpenAI is only one customer—albeit a high-priority one—for Microsoft Azure. Microsoft still needs to provide critical cloud infrastruture for government, financial services and banking, healthcare, manufacturing and more.
Microsoft has been working closely with OpenAI to build out the necessary infrastructure needed to support the continued advancement of their AI models, with plans for a data center project that could cost up to $100 billion. This project includes the development of an AI supercomputer called "Stargate," set to launch in 2028. However, even with these ambitious plans, it has become clear that the demand for AI is outpacing the rate at which infrastructure can be built.
Oracle is looking to capitalize on this opportunity by offfering up its cloud infrastructure to support AI workloads. Over the last year, the company has secured valuable partnerships with X, Microsoft and Google Cloud. Now, a new partnership between OpenAI, Microsoft, and Oracle will also allow OpenAI to leverage OCI to extend the Microsoft Azure AI platform, providing much-needed capacity for OpenAI's services. With these strategic alliances, Oracle is not only boosting its own business but also solidifying its role as a pivotal player in the AI infrastructure landscape.
Importantly, this partnership does not change OpenAI's strategic cloud relationship with Microsoft. All pre-training of OpenAI's frontier models will continue to take place exclusively on supercomputers built in partnership with Microsoft. The collaboration with Oracle simply allows OpenAI to use the Azure AI platform on OCI for inference and other needs.
As the demand for AI soars, cloud giants and investors are gearing up for what could be the largest CapEx the industry has ever seen. While profits may take a backseat in the short term amidst this frenzied scramble for compute power, one company has already emerged as an early winner: NVIDIA. The company boasts an estimated 80% market share in data centers, and its data center business, which includes GPU sales, has experienced explosive growth. In the most recent quarter, NVIDIA's data center revenue surged to $22.6 billion, accounting for a staggering 86% of the company's total sales.
The market has taken notice of NVIDIA's unrivaled position in the AI chip market, propelling the company's market capitalization to a staggering $3.09 trillion last week, surpassing even Apple. This represents a breathtaking 218% increase since last year.