EMAIL DETAILS

Team, here are my notes from the technology DD call with Roboteam this morning.

Neil, thanks for all the hard work of going to Tel Aviv and getting us this far with Roboteam.

-------

Engineering team:
25 engineers, 6 mechanical engineers, 7 software engineers, 8 HW embedded systems engineers, 4 “Innovation Lab” engineers for general prototyping. UI is outsourced to 3 engineers.

Patents:
They have 8 patents filed and pending.

History:
The team started 4 years ago leveraging a public domain robotics platform called, Aria. http://robots.mobilerobots.com/wiki/ARIA
Since then they have developed their own robotics platform.

Their first project was called Sapphire. It was for capturing images with 2 short range cameras and 2 long range cameras. Sapphire is no longer used because autonomous features is not where the company is focusing on. The current Laser sensor is an off-the-shelf components they buy for image mapping.

What levels are Roboteam autonomous?
Autonomous navigation from point A to point B is not implemented yet.
Autonomous 1) If MTGR finds out it is climbing steps, it knows how to climb better. This helps the operator to control the robots at a greater ease.
Autonomous 2) If MTGR flips, it knows how to right itself without the operator having to know how to remotely manipulate the arms to do so.
For this military market, customers are not looking for autonomous features. Customers want operator control.

Technology Focus:
The company’s focus is mainly on Remove Control capabilities with cameras to provide situational awareness to a team of operators in a highly volatile networking environment.

Building Blocks:
C++ and C# are the main languages behind their core technology. Java and WPF (Windows) are used for UI layer. Java on Android and WPF on Windows. MTGR uses Embedded Linux which is also quite standard. IRIS uses custom micro-controllers.

A Reliable Mesh Network in a Highly Unreliable Networking Environment:
Their Mesh Network architecture diagram is fairly straight forward with several standard OSI like layers. This allows them to create a SDK (Software Development Kit) that 3rd party developers can use to build Mesh Networks. I believe their Mesh Network SDK is currently used internally, more as a software development framework.

Technologies that are used are fairly standard as they are not trying to reinvent the wheel. For example, Reliable Multicast Protocols had been around since the 80’s. There are many algorithms that they can choose from to implement their Mesh network. Cassandra is 3rd party database that they use to provide “eventual consistency” across databases on a serverless network. The harder problem they are solving is to provide a fast reliable Mesh network in a highly unreliable networking environment. I would say this is the key on their wins.

In the early 2000 during the early days of usage of SUGV (Small Unmanned Ground Vehicles) in Iraq and Afghanistan , these robots are mainly used in an individual bases. After a decade, the usage of these robots is becoming more common. They are now used in more of a group settings where it is not uncommon to see 20 operators with 20 robots in one environment operating concurrently as a team to solve a problem or achieve a mission.

Yosi claims their competitors cannot stream videos reliably. This is the key to operate the robots effective. Without reliable videos, the operators are blind.

Competitive Advantage:
MTGR is cheaper, more reliable, and half of the weight comparing to their PackBot and Dragon Runner. MTGR is considered the best by the USG’s CTTSO.

Lack of Biz Dev & Government Relationships:
Yosi claims that they have not lost in any competitive situation. If they ever lose, it would be because of their lack of long term relationships with the clients comparing to their competitors which was funded by the USG for decades.

Market Opportunity:
Yosi claims there are currently about 10,000 robots deployed. iRobot sold about $500m worth of robots in the last decade. These robots are old designs and they underperform MTGR in the new way of usage of robots. They will get replaced and more robots will be used.

Pros:
+ Smart, scrappy engineers who are making these decade old robots better and cheaper.
+ Not reinventing the wheel, using lots of standards based technologies. But engineering it better for higher adoption.
+ A market that is getting more mature and proven to have high demand. Can leverage lessons learned from iRobot and Dragon Runner.

Cons:
- Lack of relationship to sell into government comparing to competitors
- Long sales cycle, government contracts can disappear anytime.

Next Steps:
* We should do some customer due-diligence.
* Study their deal pipeline to see how and whether our biz dev team can help.
* Understand why there is a big haircut to their top line.

PS (Some BS...):
About 20 years ago, one of my Masters project at Cornell was to build a Reliable fault-tolerant distributed system for multimedia encoding. I also created a server-less app (aka Mesh Network) in a distributed environment, I had to defend it in front of the current CTO of Amazon, Werner. Werner was a research scientist for Prof. Ken Birman at the Cornell systems lab. Birman founded a company called Isis which build a SDK for fault tolerance distributed system which was acquired by Stratus for $25m back in the days. Birman invested in my first company, the founding team was from this lab, which generated a 100x return.


William Lee, Managing Director
Rosemont Seneca Technology Partners (RSTP)
333 Bush Street, 21st Floor
San Francisco, California 94104
Office: 415.767.1289
Mobile: 617.407.5958
www.rstp.com

This email communication is privileged and confidential and is intended only for the individuals or entities named above. Any unauthorized dissemination of any of the contents of this email is strictly prohibited. If you are not the intended recipient, please do not read, copy, use or disclose to others the contents of this communication. Please notify the sender that you have received this e-mail in error and then delete the e-mail.