US-based technology giant Microsoft has recently focused mostly on artificial intelligence. The company recently signed a “multi-million-dollar” investment agreement with OpenAI, which developed ChatGPT and DALL-E, the most popular artificial intelligence tools of recent times. The company has also started integrating OpenAI’s artificial intelligence tools into Bing and office applications.
But OpenAI’s history with Microsoft goes back much further. The company invested $1 billion in OpenAI in 2019. In addition, Microsoft engineers would produce a supercomputer for OpenAI to continue its work. This supercomputer to be used to train artificial intelligence was far above anything Microsoft had in those years. Moreover, Microsoft did not know how to do this without crashing its cloud system Azure.
The supercomputer cost hundreds of millions of dollars!
Microsoft somehow managed to produce this supercomputer. Using tens of thousands of artificial intelligence-focused A100 graphics chips released by NVIDIA to do this, Microsoft spent hundreds of millions of dollars, according to Scott Guthrie, one of the top executives in artificial intelligence and cloud systems. Nidhi Chappell, responsible for Microsoft products, said that ChatGPT came to this day with this supercomputer.
Microsoft actually invested in itself with this supercomputer!
Microsoft also used the supercomputer it developed for OpenAI in its own projects. In fact, services to be marketed were created with the work done on the supercomputer. In other words, Microsoft sold artificial intelligence tools developed with supercomputers to its own customers.
Microsoft has already started working on the next generation supercomputers!
The supercomputer Microsoft developed for OpenAI is of great importance for the company. In fact, Microsoft has already started working on the next-generation version of this supercomputer. The new supercomputer, which will be more powerful than the existing supercomputer, will enable the development of much better artificial intelligence tools such as ChatGPT and DALL-E, which are in use today, with its faster processing capacity, in a much shorter time.
According to Scott Guthrie, Microsoft is working to solve all ecosystem problems that will not only improve the speed between servers, but also improve efficiency and cost, not just more powerful supercomputers. “The artificial intelligence that is astounding the world right now is based on the supercomputer we started building a few years ago. The new models will be built on the new supercomputer we are currently training, which will enable even greater and more complexity.” Guthrie said, in fact, much more impressive artificial intelligence tools are on the way.
40887
