Microsoft Corp. is reportedly diversifying its chip suppliers for its Azure cloud infrastructure, using highly specialized semiconductor technology from Xilinx Inc. in addition to those from its chief supplier Intel Corp.

Microsoft’s move comes at a time when it’s trying to promote the artificial intelligence and machine learning capabilities of its Azure platform, which is said to be one of the most promising growth areas in the public cloud.

According to a report in Bloomberg Tuesday, Microsoft has agreed a deal with Xilinx wherein that company’s chips will account for half of the coprocessors currently used with Azure servers to handle machine learning tasks. Previously, such workloads were powered exclusively by chips from Intel’s Altera division. Intel had acquired Altera Corp. back in 2015 in order to add programmable chips similar to Xilinx’s to its lineup of regular central processing units.

Bloomberg reported that Microsoft will continue to buy chips from Intel for its other cloud infrastructure offerings. “There has been no change of sourcing for existing infrastructure offerings,” according to a spokesperson for Microsoft.

Microsoft is said to be treading carefully with its new supplier though, as Bloomberg reported that Xilinx’s chips will need to meet performance goals before being more widely deployed.

Flexible chips such as Xilinx’s field programmable gate arrays, which are processors used in data centers that can be reprogrammed in real time for different computing tasks, are growing in popularity. Cloud providers such as Microsoft are increasingly relying on the chips, which can process AI workloads much more quickly than standard CPUs. Xilinx’s FPGAs have actually been around since the late 1980s, but it’s only with the emergence of AI that their market scope has grown.

“Programmable chips are key to the success of infrastructure-as-a-service providers as they allow them to utilize existing CPU capacity better,” Holger Mueller, an analyst with Constellation Research Inc., told SiliconANGLE. “They’re also key enablers for next-generation application technologies like machine learning and artificial intelligence.”

Despite this, Mueller said Microsoft’s decision to use Xilinx’s chips was probably more down to “good purchasing practices” as opposed to its product being better than Intel’s.

“Still, its a shot in front of the bow of the Intel ship. But no partnerships last forever, and partners need to keep proving themselves.”

Xilinx is certainly trying its best to do that, however. For example, it made headlines last month with the announcement of a new breed of computer chips designed specifically for AI inference, which relates to the application of deep learning models in consumer and cloud environments. The new Versal chips are said to combine FPGAs with two higher-performance Arm processors, plus a dedicated AI compute engine that enables higher throughput, lower latency and greater power efficiency than existing hardware.

Analysts said that with the new Versal chips, Xilinx is targeting Nvidia Corp. and its graphics processing units that have become the standard for AI workloads. In any case, it’s unlikely that Microsoft is using these at this time since Xilinx said the Versal chips won’t be released until after summer 2019.

Image: bsdrouin/Pixabay