Cost

  1. GPU server cost USD6000.
  2. Previous AI IC cost USD 90+

Energy

High energy consumption computing is not worked for Outdoor /Portable application.

4G/5G Transmission

  1. 4G/5G connection quality is still unstable.
  2. Massive transmission costs are very expensive.

AI Algorithm update

  1. Frequently AI algorithm update.
  2. Often require parallel execution of 2+ algorithms in a system, but previous solution only could run one simple algorithm one time.

Cost

  1. Require AI miniaturization engine in FPGA or ASIC.
  2. Low cost AI IC solution.

Energy Efficiency

Low power consumption AI IC solution for AI Portable units & Outdoor application.

Transmission

  1. Edge AI unit work with Cloud server.
  2. Only few data need to be uploaded to the cloud for advance analysis. Much to reduce transmissionrisk and costs.

AI Algorithm update

  1. Requires a flexible, reconfigurable architecture hardware platform.
  2. multiple algorithms can run in parallel.

 

IN A VIEW OF THIS

Due to AI algorithm just like a human brain , very huge and complicated, DeepMentor use Proprietary /Patent computing to miniaturize the whole AI algorithms with 90% off data amount & computation, then generate “the Circuit” AI IP for Customer (IC design house).

  • AI Model Miniaturization DNN Engine.
  • Flexible platform, be able to handle multiple Algorithms at the same time.
  • It’s for 40-90nm IC production process which production cost is 1000 times off 7nm.
  • Due to mature & stable 40-90 IC production, it causes the AI IC system lower power consumption, much better heat dissipation, suit for AIoT applications.
  • Proprietary AI Automatic system design tool for 3 weeks development process.