In the current era where AI is rapidly gaining popularity, Microsoft is positioning itself strongly in the field by both investing in OpenAI and entering into a strategic partnership with cloud computing start-up, CoreWeave.
This move, potentially worth billions of dollars, underscores Microsoft’s commitment to meeting growing AI needs. The article also highlights the important role of Nvidia, whose advanced GPUs are crucial in this fast-growing sector.
Artificial intelligence on the rise
Microsoft’s hefty investment in OpenAI has propelled the company into the spotlight of the AI hype. But this is not the only place where the software giant is spending money to meet the increasing demand for AI-driven services.
According to sources close to the case, Microsoft has entered into a multi-year deal with the startup CoreWeave, potentially spending billions of dollars on cloud computing infrastructure. This news comes shortly after CoreWeave’s recent announcement that it raised $200 million. The fundraising took place just a month after the company reached a valuation of $2 billion.
CoreWeave provides easy access to Nvidia’s graphics processing units (GPUs), which are considered the best on the market for running AI models. Microsoft struck a deal with CoreWeave earlier this year to ensure that OpenAI, the operator of the popular ChatGPT chatbot, has sufficient computing power in the future. OpenAI relies on Microsoft’s Azure cloud infrastructure for its computing needs, according to a source who asked to remain anonymous.
The race to generative AI started late last year after OpenAI introduced ChatGPT to the public. This technology showed that AI is able to process human input and generate sophisticated responses. Since then, numerous companies, including Google, have incorporated generative AI into their products. Microsoft has also introduced chatbots in its own services, such as Bing and Windows.
Given the growing demand for its infrastructure, Microsoft needs more capabilities to access Nvidia’s GPUs. CoreWeave CEO Michael Intrator declined to comment last month on the deal with Microsoft, but did say that CoreWeave’s revenues increased significantly between 2022 and 2023.
Nvidia’s success and the role of data centers
This year, Nvidia has shown impressive growth, with its share price rising by as much as 170%. Recently, the company’s market capitalization even briefly reached the $1 trillion mark. This strong performance was further emphasized by Nvidia’s recent quarterly forecast for July, which came in more than 50% higher than Wall Street’s estimates.
According to Colette Kress, Nvidia’s CFO, the chipmaker’s future growth will be primarily driven by data centers. This is mainly due to the increasing demand for generative AI and large language models. An example of such a large language model is OpenAI’s GPT-4, which has been trained on extensive online data using Nvidia’s GPUs. This model is at the heart of the popular ChatGPT chatbot.
It’s also worth noting that both Kress and Jensen Huang, the CEO of Nvidia, recently mentioned CoreWeave by name. Kress did this in a recent talk, while Huang cited CoreWeave in his presentation at Nvidia’s GTC conference in March this year.
Affordable computing power
According to CoreWeave’s website, the company can offer compute capacity as much as 80% cheaper than traditional cloud providers. This means developers can access Nvidia’s A100 GPUs, normally only found through the clouds of Amazon, Google and Microsoft, but at a much lower price.
In addition to the A100 GPUs, CoreWeave also has the cheaper Nvidia A40 GPUs available. These are primarily for visual computing, while the A100 focuses on AI, data analytics, and high-performance computing. Michael Intrator, the CEO of CoreWeave, indicated that some customers are struggling to get sufficient GPU capacity with larger cloud services.
He emphasized that CoreWeave can deliver these powerful services at a very cost-effective price. Meanwhile, Microsoft is in talks with Oracle about a possible collaboration. The idea is that the two companies can rent server capacity from each other when additional capacity is needed.