Intel Unveils Integrated Strategy for Implementing Cutting-Edge Artificial Intelligence (AI)
Providing a broad portfolio encompassing data centers, IoT devices, and software

Intel announced an integrated strategy that includes new products, technologies, and investment plans to support the expanded use and accelerated growth of Artificial Intelligence (AI) across the entire spectrum, from devices to data centers.

Intel is focusing on how AI is completely changing not only how businesses operate but also how people interact with the world, and is developing the industry's most extensive set of technology options to leverage AI technology across all fields, ranging from smart factories to drones, sports, counterfeit detection, and autonomous vehicles.

At an industry event hosted by CEO Brian Krzanich, Intel explained that a broader range of technologies is required to fully understand the complexity of AI and meet future expectations, and that an ecosystem capable of expanding beyond the early adopter base must be built. Furthermore, CEO Krzanich expressed confidence that Intel possesses the assets and know-how to lead computing innovation in the current landscape, where algorithms are becoming increasingly complex and required data sets are growing.

CEO Krzanich stated on Intel’s official blog, “Intel possesses differentiated capabilities in accelerating and realizing the potential of AI.” "Intel is committed to developing AI technology and is making massive investments in technology and developer resources to advance AI technology for the industry and society," he said.

Intel announced plans to introduce the Intel® Nervana™ platform portfolio, the industry's most comprehensive AI portfolio. Developed for speed and ease of use, the Intel Nervana portfolio serves as the foundation for highly optimized AI solutions, enabling more data professionals to solve large-scale global problems involving industry-standard technologies.

With 97% of data center servers currently running AI workloads based on Intel processors, Intel provides the most flexible yet performance-optimized portfolio of solutions for AI. This includes Intel Xeon and Intel Xeon Phi processors, as well as accelerators optimized for a wider range of workloads, such as Field-Programmable Gate Arrays (FPGAs) and technological innovations acquired through the acquisition of Nirvana.

Intel Unveils Nirvana Platform, a Comprehensive Data Center Computing Portfolio Dedicated to AI

Intel also disclosed details regarding the integration of Nirvana's groundbreaking technology into its product roadmap. Intel plans to test the first silicon (codename Lake Crest) in the first half of 2017 and make it available to key customers by the end of the year. Additionally, Intel announced the addition of a new product (codename Knights Crest) to its roadmap that closely integrates Nirvana's technology with its best-in-class Intel® Xeon® processors. Lake Crest is a product specifically optimized for neural networks that provides the best performance for deep learning and offers unprecedented computing processing capacity through high-bandwidth interconnects.

Diane Bryant, Senior Vice President and General Manager of Intel’s Data Center Group, said, “We expect the Intel Nirvana platform to dramatically reduce the time required to train complex neural networks and deliver innovative performance,” adding, “Intel will achieve a 100x performance improvement by 2020 to accelerate the pace of innovation in the emerging field of deep learning.”

Vice President Bryant also announced that the next-generation Intel Xeon Phi processor (codename Knights Mill), scheduled for release in 2017, will deliver four times the performance of the previous generation. He added that Intel has begun supplying preliminary versions of its next-generation Xeon processor (codename Skylake) to major cloud service providers. The new next-generation Xeon processor family will significantly enhance inference performance for machine learning through AVX-512, an integrated acceleration technology. Further performance and product specifications for this platform are scheduled to be revealed around mid-2017, the time of product launch, and will meet the needs of customers across various industries.

Saffron Technology Supports Intelligent Local Analytics Across IoT

In addition to silicon, Intel introduced leading AI solution assets for customers requiring business insights, including Intel Saffron Technology. Saffron's platform is based on memory-based inference technology and accurate analysis of heterogeneous data. This technology is particularly suitable for small devices, enabling intelligent local analytics across the IoT and helping to implement cutting-edge collaborative AI.

Intel provides general intelligent APIs that can be extended across its distributed processor product portfolio, spanning from edge devices to the cloud, to enable AI to be deployed more easily in various fields. In addition, it features embedded technologies such as Intel RealSense cameras and Movidius Visual Processing Units (VPUs).

In addition, Intel announced a strategic partnership with Google and stated that it will provide multi-cloud infrastructure that ensures openness, flexibility, and security through enterprise IT technologies. The collaboration between the two companies is expected to focus on technology integration centered on Kubernetes (containers), machine learning, security, and IoT.