Analyst Take: Continuing the expansion into AI chips is important for Intel. The company is anticipating its AI solutions to generate over 3.5 Billion USD in 2019 and of course are expecting continued growth into the next year as industries continue to make significant investment in AI.
What Was Announced:
Intel announced current and future availability of a series of new products designed to accelerate AI system development and deployment from cloud to edge. First, the company demonstrated its long awaited Intel® Nervana™ Neural Network Processors (NNP) for training (NNP-T1000) and inference (NNP-I1000) — the NNP series represents Intel’s first purpose-built ASICs for complex deep learning. The company designed these around delivering scale and efficiency for cloud and data center customers. The NNP’s have been widely talked about since the acquisition of Nervana, however, it is good to see them announced as GA.
Intel also revealed its next-generation Intel® Movidius™ Vision Processing Unit (VPU) for edge media, computer vision and inference applications. The company is making some strong claims around the efficiency of Movidius VPU versus NVIDIA Jetson AGX Xavier, however those claims will need to be tested in the real world to validate and gain market acceptance.
How Important Are These Announcements to Intel?
Let’s just say, Very Important! The announcements made by Intel were important for more than just revenue purposes. Despite the early success that Intel has had in AI, the company has been trying to keep pace with NVIDIA and the launch of the Nervana NNPs (ASICs) and Movidius VPUs have been anticipated for sometime, however with the announcement of general availability being made for certain products as well as 1H 2020 for the VPUs, enterprise, government and hyperscalers can plan and procure for deployments of Intel’s new AI solutions in the coming year.
What People Should Be Paying Attention To with AI at Intel
The short answer, is a lot. Intel’s Xeon Scalable still has a massive marketshare for datacenter compute >90% and therefore most of the world is doing at least some of their AI workloads on Intel. The company is also competing soundly for edge workloads and continues to build challenger technology for areas like computer vision for autonomous vehicles. This will place Intel squarely as part of the conversation and the company will challenge for market share in training and inference as the battle wages on to determine the best frameworks and hardware for developing and deploying enterprise AI.
Furthermore, Intel’s full stack approach that includes compute, storage, memory, connectivity and software including a massive partner ecosystem all position the company well to compete for AI training and inference workloads. This isn’t to say that the competition for Intel won’t be fierce, because it will. But these announcements should provide the company and its stakeholders some encouragement that progress is being made and Intel won’t miss the boat on the AI opportunity.
Futurum Research provides industry research and analysis. These columns are for educational purposes only and should not be considered in any way investment advice.
Read more analysis from Futurum Research:
The original version of this article was first published on Futurum Research.
Latest posts by Daniel Newman (see all)
- Cloudera Q3 Shows Momentum for its Data Platform - December 10, 2019
- Encouraging Effective and Consistent Collaboration - December 10, 2019
- Dell Tech’s PowerOne Approach to Hybrid Cloud - December 9, 2019