Microsoft has bet a good part of its future on advances in Artificial Intelligence. But not only for the company, but also so that OpenAI can advance research and development on models and related technologies so that it can give you the AI muscle you need. For this they have made a large direct investment in it, but also another very important indirect one: in the startup dedicated to the development of infrastructure for cloud computing CoreWave.
According to CNBC, Microsoft is going to invest several billion dollars in this company over several years. CoreWave, which is valued at about $2 billion and has recently closed a round of financing in which it has raised $200 million.
Among other things, it offers simplified access to Nvidia GPUs, which are considered to be the best on the market today for running Artificial Intelligence models. It is what would have led Redmond to the agreement, which ensures that OpenAI will have the computing power it needs from now on.
Currently, OpenAI depends on the Azure infrastructure. And in a framework in which the demand and supply of AI and Artificial Intelligence services will not stop creating and increasing, the demand for the infrastructure that allows you to run your workloads, train models and develop them, and create services with AI, it is going to grow exponentially in the coming years. Therefore, Microsoft needs to have as much access as it can to Nvidia’s GPUs.
He CoreWave CEO Michael Intratorhas not wanted to make statements about this agreement, although it has acknowledged that its income has multiplied more than notably only from 2022 to 2023. Precisely, due to this increase in demand for infrastructure for AI.
The financing announced a few days ago by the company, founded in 2017 and which has 160 employees, is an extension of its April round, in which it raised 221 million. In a previous one, Nvidia had invested 100 million dollars
CoreWave ensures on its website that can offer computing power at a price 80% lower than traditional cloud providers. Among other things, it offers Nvidia’s A100 GPUs, which are also offered by AWS, Google Cloud and Microsoft Azure.
Nvidia’s cheaper A40 GPUs, traditionally focused on visual computing, are also available. The A100 are designed for more demanding tasks, such as Artificial Intelligence, high-performance computing or data analytics. But for many cases, CoreWave recommends the A40, since “they do a good job with a good cost-effectiveness ratio«.
