Qualcomm: Can it spur an ARM server surge?

Qualcomm is planning to move into the ARM server market and is already working with customers. The effort will take time to really pay off, but may at least create more enthusiasm for microservers.

The initial enthusiasm about ARM-based servers and so-called microservers has fallen with a big thud. Amazon Web Services has noted that the ARM ecosystem can't keep up with Intel's server innovation. Microservers clearly aren't living up to the advanced billing.

Special Feature

Microservers: A Data Center Revolution?

The rise of cloud computing has put new and different demands on data center capacity. Powered by smaller chips like the ones in mobile devices, microservers have emerged to help meet the needs of 'scaling out' and to produce servers that are lower cost and lower power.

Read More

Perhaps Qualcomm can change that equation. On Wednesday, Qualcomm CEO Steven Mollenkopf said the company plans to enter the server market. At Qualcomm's annual analyst meeting, Mollenkopf said:

If you look at what's happening in the data center, one of the big things that is impacting the industry is the cloud. We are obviously participating in the cloud in the form of the smartphone. But if you look at the implications of the cloud in the data center it opens up the opportunity for someone like us.

Microservers: Not living up to the hype?  |  Windows Server with ARM? Crazier ideas have happened  |  HP launches 'first enterprise-class' ARM-based server

Mollenkopf said the data center has largely been proprietary and x86-based, but is opening up. Scale will matter and cloud vendors are designing their own architectures. Qualcomm's chips are evolving to work in the data center and address a $15 billion total addressable market in 2020. "We think we're uniquely positioned to go after it because of our scale in mobile," said Mollenkopf.

When later questioned about the ARM server market, Mollenkopf said the company is "actively engaged with customers" on servers. But customer engagement to meaningful revenue will take time and may not equate to material sales until 2020.

The rationale via Qualcomm's presentation goes like this:


How will Qualcomm fare in the server market? Here are a few thoughts:

  • The problem with the ARM-server market to date is that it has focused on specialized workloads. As Nick Heath noted this week, workloads change. A microserver has to be more general purpose. Qualcomm has the intellectual property and scale to potentially create a more general purpose server that can handle different workloads.
  • Qualcomm is one of the big dogs in the ARM ecosystem. ARM servers are supported by Applied Micro Devices and AMD, but Qualcomm brings scale that few vendors can match. At the very least, Qualcomm's interest is a sign that the timing for ARM servers is off, but the potential isn't. Should Samsung get involved with ARM servers, all the big guns would be aimed at Intel.
  • Mollenkopf was smart to note that the ARM server move would take time to deliver results. The ARM server market will have to expand beyond cloud providers to really scale. Qualcomm has the capital and research and development budget to be patient.