TECH NEWS – According to Reuters, the Redmond-based tech giant and the company behind ChatGPT (in which Microsoft has invested!) would jointly develop hardware…

The artificial intelligence industry is slowly moving towards proprietary solutions, and now we hear that Microsoft and OpenAI are developing their own chips to compete with Nvidia’s H100 chip. Nowadays, the “greens” have few rivals in the hardware market, and so many people are ordering from them (which in turn has driven up the price of traditional desktop and workstation graphics cards…), and it’s no joke to say that the H100 is selling like hotcakes, and Nvidia is making a huge profit from it.

But because of the high demand due to the backlog of orders, the increased costs and the relationship with Nvidia, “exclusive” access is happening nowadays. So it’s not exactly a balanced situation between the companies (those who are on good terms with them get hardware, those who aren’t don’t), and Microsoft may be preparing to upset the status quo as they believe the AI chip industry is hindering their goal (generative AI resources that can get into mainstream programs that many of us use).

Microsoft has codenamed the chip Athena, about which not much is known. It’s due to be unveiled at the company’s Ignite conference in November and would like to be on par with Nvidia’s H100, but that won’t be easy as developers prefer Jensen Huang’s chip because of its CUDA platform and good software. Catching up in terms of computing power is not difficult, but it will take time to develop the software resources. So far, OpenAI has been using Nvidia’s A100/H100 pair to power ChatGPT, but that may change in the future, but it’s safe to say that Microsoft’s solution will be in its infancy for a while.

Let’s not forget that the GeForce RTX 3000/4000 cards also have tensor cores that can be used for AI. So AI is already starting to take off in that sense, so it’s easier to use the technology for things like Photoshoppers (for custom backgrounds).

Source: WCCFTech