Edge-native AI and New Opportunities for Semiconductor OEMs
Automotive and device OEMs face challenges associated with the cost and latency of cloud processing. MicroAI™ brings AI-enabled telematics to the edge.
Smart Automotive
23913
post-template-default,single,single-post,postid-23913,single-format-standard,bridge-core-1.0.4,translatepress-en_US,bridge,ajax_fade,page_not_loaded,,qode_grid_1300,qode-content-sidebar-responsive,qode-theme-ver-18.0.6,qode-theme-onetech,disabled_footer_top,wpb-js-composer js-comp-ver-5.7,vc_responsive
 

Edge-native AI Deepens the Semiconductor Value Chain

08 Oct 21
Edge native

Edge-native AI Deepens the Semiconductor Value Chain

The global artificial intelligence (AI) market is expanding at an accelerated rate. Historical and future market sizes can be a bit tricky to calculate, but there are a couple of generally held estimates: the global market was ~ $30 billion in 2018; the market will explode to ~ $250 billion by 2025.

This phenomenal growth in the AI market will create new opportunities for the semiconductor industry to deepen its penetration into the technology value chain. But, there will be challenges. The semiconductor sector will need to leverage new AI-enablement solutions to maximize the value of their products within a constantly evolving market.

The Industry Challenges

The rapid growth in AI applications across virtually every industry sector (automotive, medical, telecom, energy, etc.) has not yet created equivalent growth for chip manufacturers. Challenges have included:

  • Development of integrated solutions for the IoT/IIoT market: Utilization of AI technology to develop new chip-driven solutions within existing and emerging IoT/IIoT markets.
  • Limited insights into chip performance: Chip OEM’s lack critical insights into the functionality and performance of chips running on smart devices and machines.
  • Capturing more value from the technology stack: Adoption of Edge and Endpoint AI solutions to generate more value within the IoT/AI technology stack. Semiconductor companies currently have a market capture rate of only 15 to 20 percent.
  • Creation of new micro-verticals for AI-enabled chip products: The use of embedded AI technology to develop new end-to-end solutions for targeted verticals and new micro-verticals.

 

The Solution

Traditional Edge-based AI solutions will not meet the above challenges. Legacy solutions simply cannot provide the semiconductor industry with the chip-centric approach to AI and machine learning (ML) necessary to move their products deeper into the value chain. The industry needs an approach that is truly at the “edge”.

Edge-native AIintelligence at the extreme edge—will bring the benefits of AI and ML much closer to the semiconductor manufacturer. Edge-native AI delivers chip-level intelligence that will enable:

  • AI-enabled on-chip memory: Embedding nonvolatile memory at the chip level will augment the performance of AI algorithms.
  • Development of higher-speed hardware: Chip-centric, Edge-native AI, will drive improvements in hardware processing speed.
  • MCU-level training of smart assets and machines: AI embedded directly onto the MCU of an asset will allow chip OEMs to create new micro-verticals.
  • Deeper visibility into chip performance: Chip manufactures will have the ability to track the application and performance of their products.
  • Improved semiconductor ecosystem cyber security: AI at the extreme edge improves cyber-security by eliminating the need to route chip/asset data to the cloud for processing.

 

Edge native

 

The Impact

Edge-native AI will provide a true paradigm shift in AI-enabled capabilities for the semiconductor industry. Tangible impacts will include:

  • Differentiated offerings: Being able to leverage chip-centric AI and ML will allow semiconductor companies to develop new offerings tailored to the specific needs of a customer or industry.
  • Improved chip performance: Embedded intelligence will drive the acceleration of on chip memory while allowing OEMs to actively monitor chip data and performance.
  • Increased revenue capture: Semiconductor companies will be able to significantly expand their market capture rate within the overall technology stack.
  • Improved security: Non-cloud-based AI will improve cyber security throughout the semiconductor ecosystem (chip manufactures, hardware manufactures, asset OEMs, and end-users).