Microsoft builds its own AI hardware: Project Brainwave

Microsoft has teased that it's working on AI technology, with the heart and soul of Redmond's artificial intelligence powered by Intel's new Stratix 10 field programmable gate array (FPGA) chip.

Microsoft calls its AI hardware platform "Project Brainwave", which is capable of a huge 39.5 teraflops of machine learning performance, with less than 1ms of latency. What makes Microsoft's ventures into AI better than their competitors? Well, the new chip handles complex AI tasks as they're received instead of batching the operations together and working on them after.

Thanks to Project Brainwave relying on a "soft" dynamic neural network processing engine using off-the-shelf FPGA-based processors, Microsoft's own AI framework (Cognitive Toolkit) also works with Google's own TensorFlor, and other systems. This means you can build a machine learning system, and run everything in real-time instead of being bottlenecked by hardware.

Microsoft wants to push out its new Skynet Project Brainwave systems into their Azure cloud services, so that companies can use super-powerful, live AI. We don't know what Microsoft will achieve from this in the short term, but you can bet your bottom dollar this will turn into something big in the next 5-10 years, and beyond.