Periodic Reporting for period 1 - Reexen (Ultra-low cost & ultra-high efficiency AI processor for enabling fast and cost-effective deployment of edge-computing applications)
Berichtszeitraum: 2020-01-01 bis 2020-04-30
In the edge AI applications for video surveillance, smart city & office, wearables and industrial smart manufacture there is a growing need for high throughput & computing efficiency at low power; there are concerns about privacy & security and pressure to reduce size & cost.
The society faces a challenge to bring AI to the edge in order to leverage it to the maximum without the need to upload data to the cloud (security concerns) requiring high energy consumption and high bandwidth. Related to Covid19 pandemic there are emerging edge AI developments e.g. for temperature measurement and monitoring people in a space.
Reexen’s accelerator chip allows the sustainable implementation of smart urban mobility (smart cities/office) and mobilizing the industry to adopt more resource efficient and clean manufacturing processes.
The main objective of the project is to develop a ultra-low power high throughput AI accelerator chip that is optimized for wearables, smart industrial robots, video surveillance/ smart city and smart office applications on the edge.
The society faces a challenge to bring AI to the edge in order to leverage it to the maximum without the need to upload data to the cloud (security concerns) requiring high energy consumption and high bandwidth. Related to Covid19 pandemic there are emerging edge AI developments e.g. for temperature measurement and monitoring people in a space.
Reexen’s accelerator chip allows the sustainable implementation of smart urban mobility (smart cities/office) and mobilizing the industry to adopt more resource efficient and clean manufacturing processes.
The main objective of the project is to develop a ultra-low power high throughput AI accelerator chip that is optimized for wearables, smart industrial robots, video surveillance/ smart city and smart office applications on the edge.
1) We defined a technical roadmap to finalise the specifications and features of the commercial version; defined in-field pilots to further optimize product development and validate the suitability of our solution. Carried out risk analysis for all technical activities indicating the mitigation and contingency measures.
2) We conducted market study on AI devices, services and products with edge-computing needs, re-defined a commercialization and dissemination plan together with a partnering assessment to guarantee successful scale-up and benchmarked key customers in each vertical and strategically redefined our target verticals to secure smoother market entry and expansion. We have confirmed FTO and re-defined IPR strategy.
3) We detailed resources needed to reach TRL8 and further TRL9, updated pricing strategy and financial projections and determined the expected profitability and payback period.
2) We conducted market study on AI devices, services and products with edge-computing needs, re-defined a commercialization and dissemination plan together with a partnering assessment to guarantee successful scale-up and benchmarked key customers in each vertical and strategically redefined our target verticals to secure smoother market entry and expansion. We have confirmed FTO and re-defined IPR strategy.
3) We detailed resources needed to reach TRL8 and further TRL9, updated pricing strategy and financial projections and determined the expected profitability and payback period.
Our chip goes far beyond the state of the art with its technical attributes of efficiency of 100 TOPs/W and ultra-low power consumption. This gives our customers an opportunity to bring AI to the edge without compromising in computation capacity while having the industry’s lowest energy consumption allowing market-wide paradigm shift towards reaching the Green Deal objectives. Apart from a high revenue and net income projections and boosting the EU edge AI industry, we will create 19 new FTE by the end of 2026.