In what can only be described as a warning to Intel, Microsoft is developing its own silicon for use in its cloud data centers and consumer focused Surface products, according to a Bloomberg report.

Microsoft is reportedly developing an Arm-based processor for its Azure public cloud offering. The news shouldn’t come as a surprise as Azure has already been deploying Marvell Arm-based ThunderX2 CPUs — alongside more traditional x86-based processors from Intel and AMD — for the better part of a year. Meanwhile, competing cloud providers, including Amazon, have also deployed or announced plans to deploy Arm-based silicon.

News of Microsoft’s reported plans to move away from Intel-based CPUs in either its public cloud or its PC offerings sent Intel’s stock price tumbling after the market closed on Friday afternoon. As of publication, the company’s stock price was down more than 6%. Meanwhile rival AMD’s stock price was down less than 1%.

The news comes amid an industry wide shift away from Intel processors in favor of AMD’s latest generation of EPYC processors and a new crop of Arm-based CPUs. Earlier this month Amazon doubled down on its commitment to developing its own silicon, announcing plans to shift more of its services over to its own Graviton2 processor.

“We have a deep relationship with Intel and AMD, and we will for the foreseeable future,” AWS CEO Andy Jassy said during the AWS re:Invent 2020 virtual conference. “However we realized a few years ago if we wanted to continue to push the envelope of price to performance we knew we were going to have to develop some of our own chips.”

Likewise Oracle announced plans to deploy Ampere’s 80-core Altra CPUs on its own public cloud offering back in September.

Bloomberg also reports that Microsoft may also be exploring another custom chip for use in the company’s Surface line of PC computers. Like in Azure, Microsoft is already using Arm-based processors from Qualcomm in at least part of its Surface lineup. The Surface Pro X, announced last year, uses Qualcomm's SQ-series chipset.

If true, Microsoft would be following Apple's footsteps. Apple, in June, announced it would begin phasing out Intel processors from its popular Macintosh product line in favor of its own custom silicon. The first of these Arm-based Macs launched in November, and a complete transition away from Intel is expect to take two years.

However, Bloomberg reports a server chip is more likely as Microsoft’s chip design unit reports to Jason Zander, head of the Azure cloud business and not to Panos Panay, who oversees Surface products.

In a statement late Friday, Intel attempted to downplay the news and implications it may have for its long-term relationship with Microsoft, arguing that emerging workloads, like AI, were "driving more silicon experimentation in the cloud." The company affirmed its commitment to developing x86 CPUs as well as new products, such as GPUs and AI chips, to meet the demands of emerging cloud workloads.

While Intel started the year on a strong note, the company has experienced repeated blows over the course of 2020.

Intel ended a string of strong quarters earlier this fall with a downturn in the third quarter that saw the chipmaker’s revenues slip 4% and its net incomes plummet nearly 30% year over year. The quarter also saw Intel’s data-centric revenues — a growth driver for the company, accounting for more than 50% of quarterly revenues — tumble.

Intel’s slipping data center revenues were contrasted against strong performance by rival AMD, which saw its revenues climb 56% to $2.8 billion during the same quarter, propelled in part by the success of its second generation EPYC server processors. The company also announced the acquisition of Xilinx, Intel's biggest competitor in the FPGA market, during the third-quarter earnings call.

However, beyond eroding confidence in Intel’s product line, the chipmaker has faced production challenges and delays both with its current generation 10-nanometer manufacturing process and with its severely delayed 7-nanometer process.

To date, Intel has yet to ship a 10-nanometer Xeon Scalable chip, and it isn’t expect to until early next year. And the company’s upcoming 7-nanometer process isn’t expected to ship until late 2022 or early 2023.

By contrast all of AMD’s current generation processors are built using Taiwan Semiconductor Manufacturing Co.’s 7-nanometer process while Apple is already using the fab’s 5-nanometer process.

TSMC’s roadmap has the chip fab launching its first 3-nanometer processors in late 2022.

Editors note: This story has been updated to include a statement provided by Intel.