Startup iryx Launches, Led by Ex-Arecont and FLIR Executives

Suddenly, the video surveillance industry is witnessing a new wave of startups.

Now launching is iryx, a startup led by veterans of 2 of the industry's first wave of megapixel camera companies, Ian Johnston, most recently VP of Engineering at FLIR, earlier CTO of IQinVision, and Raul Calderon, most recently President at Arecont Vision.

iryx describes their upcoming offerings as 'machine learning camera products that offer sensor fusion.' We spoke with the executives about the company's launch and analyze their outlook and positioning.

Offering Overview

Unlike most AI companies that have targeted cloud or add-on AI offerings, iryx is building hardware - cameras, or at least what they see as the next generation of cameras. Calderon described it as they:

utilize sensors looking at different parts of the electromagnetic spectrum such as visible, thermal and mmWave radar

The goal of multiple sensors, and in particular fusing them, is that they have different strengths. Visible imaging is ideal for humans but presents greater challenges for computer vision. Thermal imaging eliminates key problems in shadows and lighting common in visible imaging. Finally, radar provides richer information in terms of objects positioning and tracking that is hard to do with visible or thermal imaging.

Radar is currently atypical for commercial security applications but its use is expanding, driven by self-driving cars, where the price, performance, and compactness are improving. Embedding radar in cameras will become more common in the 2020s.

With those fused sensors, iryx plans to do analytics inside the device, as they describe, using a:

computer vision (CV) engine with convolutional neural network (CNN) for superior recognition, classification and identification with low power consumption at the edge.

iryx products can then send data to VMS and other management systems for response, display, action, etc.

Target Markets

The company is targeting the high end of the market - commercial, industrial, government applications where the need for more accurate security and outdoor analytics stand out.

They noted they are not trying to compete in low cost or budget applications, which would infeasible given the multiple sensors, etc. being used.

ODM / OEM and Channel Sales

The company is looking to partner with ODM and OEM partners who would incorporate this under their own brand as well as to channel (i.e., integrator) sales under the company's own brand.

Given the time and effort it takes to develop a company's own channel from scratch, ODM / OEM provides a shorter path to larger scale deployment and revenue.

iryx they will not be selling through distribution, given the higher end of the market they are targeting.

Start Shipping First Units In A Year

The company says they plan to start shipping first units in a year. We would think volume sales would be minimally early 2022, as it takes time for testing, refinement, etc. However, the company is already working on the product and has an experienced team which will cut down some of the time needed.

Not Made In China

While the company has not finalized where the product will be manufactured, they noted that it would not be made in China, given security concerns of their target market.

Johnston CEO of iryx and CTO of Cognize

One atypical aspect of iryx is that company CEO Johnston is simultaneously CTO of Cognize. Cognize is led by ex-Arcules CEO who left Arcules last summer. Johnston clarified his roles at both companies:

iryx is focused on multi-spectral sensors, combined with machine intelligence edge compute engines. It is essentially a continuation of the vision of ISD to bring innovation to the edge. ISD was primarily focused on edge storage, and recording visible light sensors. iryx expands our sensor mastery across the entire electromagnetic spectrum. All that is packaged with powerful compute engines to handle all the modern AI frameworks that our customers want to run.

Cognize is a Cognitive Awareness software solution that processes data from the world around it, giving our customers actionable intelligence, and knowledge of what is happening in their worlds. We have yet to go public with what we have in mind, but it is truly exciting.

The net result is a lot of work, but the two are highly complementary, and therefore more easy to manage.

Strengths For iryx

We see points in iryx's favor:

• Experienced executives reduce the time for both product and business development.
• Shortage of new camera manufacturers makes it easier to stand out.
• Tapping into rising trends including both radar and computer vision.

Challenges For iryx

On the other hand, there are concerns:

• How expensive will it ultimately be? The company says it will be significantly less expensive than existing multi-spectral approaches, which we believe, but it may be much more expensive than regular 'cameras'.
• How much better will their analytics / intelligence be than 'regular' cameras or generic AI on conventional video? Will their conventional rivals be 'good enough' for many applications?
• How big of a market place will there be? This relates to how much better. Will this eventually be a niche for the very high end or can it be a more broadly used product?

Outlook

It is certainly too early to tell how well they will do but it is good for users to have more options and different approaches to choose from.