The State of Industrial IoT — Notes from South Summit Panel

Earlier this month I was in Madrid for my first SouthSummit. I spoke in 3 sessions, including a panel on IoT with … 2 other Frenchmen :)

  • One was the renowned Silicon Valley investor Jeff Clavier, founder of micro-VC-turned-early-stage fund Uncork Capital (Eventbrite, Sendgrid, Fitbit, etc.).
  • The other was the Paris-based Nicolas Iordanov, Principal at Iris Capital, mostly active in Europe (France + Germany) where it deployed over $1B across over 300 companies.
  • I was representing SOSV (our fund with US$400 million under management) and HAX (the hardware branch with >200 investments).
  • The panel was moderated by Berlin-based Tobias Schirmer of Join Capital.
The panel didn’t go South

Here is the transcript slightly edited. The video is at the end of this post.


Q: What is Industrial IoT?

Definitions vary so we tried to agree on a scope.

Ben (HAX): My take is that it covers things related to manufacturing and automation. Within manufacturing:

  • Upgrading existing machines. By adding sensors and analytics.
  • New machines. With new capabilities.
  • Shrink machines. And take them from the factory to the workshop or desktop.

Jeff (Uncork): It feels like the Internet of Things has been right around the corner for … 20 years. We look at sensors and platforms. The retrofitting of machines in manufacturing lines, but also in industries like mining and so on who need the sensors and connectivity to:

  • Know what’s going on,
  • Predict downtime and maintenance,
  • Optimize to produce more with the same machines.

Nicolas (Iris): We feel the trend has two parts:

  • Edge computing — allowing local computation and analysis to happen.
  • Connectivity. It includes RFID, Bluetooth (BT / BLE), Ultra Wide Band (UWB), or broader lower-bandwidth networks like SigFox or LoRa (Long Range). Larger scale connectivity get into the manufacturers. That’s really interesting because the manufacturing process has hundreds or thousands of parameters coming together for a single output so it’s very suitable for software and A.I. analytics.
IoT is in our hands

Q: When did IoT become a thing for you?

Jeff (Uncork): It’s been around for 20 years but the adoption rate and the unit economics might now be at the point when sensors, compute power and connectivity is cheap enough to embed into new or existing machines. So it’s ready to retrofit manufacturing lines.

The problem is that it’s not an easy SaaS-like sale. It’s a heavy-duty enterprise sale with long sales cycle and massive friction to deployment. Customers want to get it done without disrupting their work, but in order to leverage IoT you have to change the way you do things, and manufacturers are very hesitant to do that.

Ben (HAX): Some companies find non-disruptive ways to upgrade machines. One of our portfolio companies, Amper, based in Chicago, uses simple, non-invasive sensor clips around the wire providing power for the machine. It collects electrical energy data to track the performance and identify underlying problems. A solution that can save hundreds of thousands of dollars caused by downtime is an easier sale.

Amper makes your factory smart in seconds

Selling new machines is more complicated. One example at HAX is Elephant Robotics, which sells cobots (collaborative robots). Cobots were heralded as the future of manufacturing: working alongside humans, saving space in factories. But so far it looks like the main benefit is to have cheaper robots to automate new tasks. Elephant installed robots at large automotive equipment maker to simply push buttons.

Note: the Boston-based cobot pioneer Rethink Robotics, which had raised almost $150 million, announced it shut down the day after our panel.

Factories often calculate “Is this going to pay for itself within 24 months?” If the answer is yes, it’s a no-brainer. Then if it works on one line, you deploy across all factories. That’s the key difference between consumer and industry: industry calculates ROI. If you can prove ROI, it sells.

HAX started 6 years ago to invest in hardware, initially mostly in consumer tech — our startups did over 100 Kickstarters — but gradually founders were less and less hobbyists, tinkerers and makers, and more and more people with industry experience, people with PhDs. These days consumer tech is less than 20% of our investments. The other 80% is industry, health and enterprise. This probably started about 3 years ago — we’re at the forefront of this change as we sign the first check on a prototype.

Nicolas (Iris): We’re quite new to IoT — less than 2 years. One example, Exotec, is a fully automated infrastructure to help retailers store and pick-up all their B2C orders. They install fleets of dozens of robots. They provided great ROI to their customers through intelligence and automation.

Ben (HAX): Is it like the Kiva robots in Amazon warehouses?

Nicolas (Iris): It’s like the next-generation, but denser as they add a third dimension and can go up to 12 meters (Kiva does 2.5m).

Skypods by Exotec deployed at CDiscount

Another company is Braincube, a co-investment with Siemens we’re announcing next week (updated here): it’s a software platform for production lines using thousands of parameters to optimize the conditions of production, already deployed on 800 lines around the world. Downtime is a cost, but so is over-production, so they help optimize use of raw material and energy consumption to save money.

Q: Who will buy these startups?

Jeff (Uncork): There is a lot of skepticism around IoT because it’s been talked about for so long. But strategics and Asian investors are willing to invest. U.S. investors are looking but think ‘I’ve seen it before, but is the time right this time?’.

It’s the same with commercial drones: everyone agrees it’s a real market opportunity but will it materialize in 3 years? 5 years? So we’re looking at companies that can get to revenue quickly and can get people interested in their series A or B.

For instance, one of our companies just closed their series A and the lead is a very large manufacturer. It is very unusual for us to bring in a strategic at series A but given that they’re going to use the technology it’s the best of both worlds.

Tobias (Join): 10 years ago the incumbent in commerce and media being disrupted were very much asleep, but it looks like industrial players are not.

Jeff (Uncork): Yes, industrial players are all looking at making investments. They have venture arms. They are very active. One of our companies just got investment from BOSCH. Another from Next47 (venture arm of Siemens).

Ben (HAX): To quote a few recent exits over $100 million:

  • 2 robotics companies from Denmark — Universal Robots and MiR (Mobile Industrial Robots) — were acquired by U.S.-based Teradyne.
  • A Japanese IoT startup was bought by Japan’s telco KDDI.
  • Germany’s Kuka, which was a public company, was bought by the Chinese appliances maker Midea for $5 billion.

But those are ‘older’ companies. The new generation is more nimble, with cheaper products suitable for more situations.

I agree with Jeff that strategics are active. Asian investors are active from either the manufacturing side, and corporates like Tencent and Alibaba also putting money into hardware. Chinese investors are less afraid of hardware as they have a giant supply chain right in their backyard, whereas in Silicon Valley almost everyone moved to software a long time ago. Going back to hardware almost feels like a step back. Some have been burnt by prior experiences that were premature compared to where we are today.

Q: How do you compare IoT startup metrics and others for investment?

Nicolas (Iris): There are opportunities today with some vertical use cases. But overall companies in the sector have longer development cycle before they are ripe enough for an acquisition. The industrial ecosystem is very mature and can write large checks but they have not proven yet they might do that only for acquiring technologies. So when we look at investments we try to project their development further than we would with other companies, to see how to get them in the hundreds of millions in revenue before looking for an exit.

Ben (HAX): To say a word about other industries: construction, mining, oil and gas, have been lagging behind in terms of innovation. We invested recently in a Canadian company named RockMass that makes a device that can measure the safety of rock structures much faster than manually or with LiDAR. Those industries are not ‘sexy’ but are worth many billions. If you can solve problems you have very little competition.

Q: Who are the founders. Is there a renaissance of the engineering entrepreneurs?

Jeff (Uncork): Some are manufacturing veterans. Others are part of this engineering renaissance.

Ben (HAX): For us it’s roughly 1/3 young grads, 1/3 people with a few years of experience, and 1/3 of ‘veterans’.

### END ###