Microsoft Intends To Develop Ai Chips On Its Own. Should Microsoft Supplier Intel Worry?

CNBC reported in June 12th that at the end of March, the Microsoft (Microsoft Corp.) Azure public cloud department issued at least three signs to announce and solicit personnel related to the development of artificial intelligence (AI) chip functions. In April of this year, the team launched the chip project manager, collaborative design of hardware and software, and AI acceleration and optimization of engineer positions.

Google (Google) executive director Sundar Pichai published the third generation TPU (Tensor Processor Unit, TPU3) at the 2018 Google I/O conference in May 8th. The new TPU can help Google improve AI speech recognition, identify specific objects in photos and videos, and understand the hidden emotions behind the text. Moor Insights Strategy analyst Patrick Moorhead estimated that Google spent $2-3 on the TPU plan.

Microsoft Bing blogger announced in March 26th that it has created a deep neural network through Intel (Intel) Arria and Stratix field programmable logic gate array (FPGA) chips, and then developed a deep learning acceleration platform called "Project Brainwave". Bing's intelligent search function is to perform large-scale machine reading comprehension with the huge computing power of this platform.

Reynette K. Au, the marketing vice president of Intel's program solution group, released a press release (pictured out) at the time that, with Intel's FPGA chip, the smart search for Microsoft Bing could understand the user's intentions and provide answers instead of providing a string of Web links that need to be handled manually.

According to the Nikkei Asian review published in November last year, the Nikkei and Elsevier analysis showed that the first place in the 2012-2016 year AI related research report, which was the most cited number, was Microsoft.