Recently, a leading technology publication reported that Microsoft was planning a data center project that would comprise state-of-the-art supercomputing ability and specialised server chips to power OpenAI’s artificial intelligence. The Information based this on inputs from three unnamed sources, one of whom had reportedly spoken to OpenAI CEO Sam Altman.
OpenAI has been the talk of tech world ever since its launch, and is a formidable contender in the ongoing battle for AI dominance. But in order to become the top player in a highly competitive AI market, any ambitious player must remember that not only does AI require extremely high compute power, it will be very expensive to put together the necessary infrastructure, given how expensive AI chips are.
Earlier this month, NVDIA CEO Jensen Huang told CNBC that the company’s AI chip, called Blackwell, would cost between US$ 30,000 to US$ 40,000 per chip! He said that they had to invent new technology to make the chip and estimated that they had invested US$ 10 Billion!
But, according to the report in The Information, Microsoft is expected to leave no stone unturned in building this project and is expected to invest a whopping US$ 100 Billion into it. The project is reportedly divided into five phases, and is currently in phase three, with phase four expected to be launched by 2026. The fifth phase has reportedly been named Stargate.
But the report did not name any of the sources cited in the story, nor has any official announcement been made by Microsoft or OpenAI so far. Watch this space for more…