Partnering with Promising Startups

Next47, Siemens’ global venture arm, is partnering with promising startups. An example is a company developing technology designed to cut the costs and energy requirements of the processors in autonomous vehicles by a factor of 1000. A look at Next47’s current portfolio illustrates how the firm is investing in innovations and partnering with leading entrepreneurs to understand the megatrends shaping the digital economy.

The future of mobility, shaped by connected e-mobility and autonomous vehicles; the future of manufacturing with 3D printing and accelerated entry of products to the market; the future of buildings and cities based on increasingly intelligent infrastructures – as a conglomerate focused on our digital future, Siemens is involved in shaping these major technological trends. In a world in which the pace of emerging innovations has tremendously accelerated and the manner in which they come about is faster, more agile and, in a certain way, more frenzied, Next47 has internalized a startup-inspired way of working in many of its areas.

 

Next47, Siemens’ global venture arm, invests in the kinds of startups that promise to create the industries of tomorrow – startups whose technologies can be magnified through the power of the Siemens ecosystem. The partnership between Next47 and its portfolio companies extends far beyond venture capital. With the support of Next47 and Siemens, startups can create new opportunities where reality sometimes glaringly lags behind possibilities. The result is an innovation pipeline that feeds Next47 and Siemens with the insights they need to stay knowledgeable about tomorrow’s shifting industries.

Autonomous Vehicles for Mass Markets

 

“Today, when you open the trunk of most prototype autonomous vehicles, you find GPUs or tensor processing hardware worth $10,000 or more,” says Forrest Iandola, CEO and co-founder of DeepScale and one of Next47’s newest entrepreneurs. “That’s more power than you need for such vehicles. It also takes up too much space, and costs too much.”Over the past several years, Iandola’s company has specialized in making precise, efficient, deep neural networks (DNNs). Such DNNs offer the intellectual horsepower behind autonomous vehicles. They are responsible for the interpretation and classification of data that these vehicles collect through sensors. And everything they do needs to happen in real-time so that vehicles can react to surrounding circumstances.

 

“We are unusual in this space because we already understand how to make deep neural nets efficient,” says Iandola. “Instead of using a datacenter-style computing platform in the trunk of a car, we are squeezing state-of-the-art DNN accuracy onto low-cost, energy-efficient processors for the automotive market.” DeepScale’s goal for driver-assistance is to run DNNs on <10W processors to fit the cost and size requirements that will allow the latest collision avoidance and convenience features to reach as many vehicles as possible. For autonomous driving, DeepScale plans to decrease the unit of measure for required processing power from kilowatts to watts, which would dramatically reduce system costs and size, while improving reliability. Providers of connected infrastructure and e-mobility such as Siemens would benefit from this step.

Leaving the Competition in the Dust

 

Verkada, a manufacturer of intelligent surveillance cameras and another Next47 portfolio company, is also confident in its ability to set itself apart from the competition. “Our competitors concentrate on camera hardware and ignore software, resulting in an experience that’s like recording to a VCR tape,” says Verkada co-founder and CEO Filip Kaliszan. Standard security cameras operate in isolation, each on its own, and recordings are only analyzed when something happens, after the fact. Verkada’s objective, on the other hand, is to allow scene analysis to be integrated and available in real-time in modern IT infrastructures. Kaliszan says this will be possible thanks to his cameras, which are software driven and intelligent.

Our competitors concentrate on camera hardware and ignore software, resulting in an experience that’s like recording to a VCR tape.

Verkada’s platform can already be used with a smartphone to search precisely for incidents, such as a door opening, without having to watch entire video recordings from multiple cameras. As Verkada continues to improve its platform, its goal is to provide software that will automatically send an alert to a smartphone in real-time with precise information about what is happening. That would enable users to take action regardless of their location.  

 

“I’ll give you a funny example of how we’ll get there,” says Kaliszan. “Each of our cameras has an accelerometer in it. We built it in to provide tamper detection and prevent interruptions or gaps in security. But we’re also located in California. Because we have accelerometers on every camera we can also detect when an earthquake happens, and then take action on that data - notifying the customer and authorities.”

 

For Siemens, a leader in intelligent building technology and infrastructures, the integration of such cameras in a comprehensive IT infrastructure would be a complementary technology for its existing palette of smart services. Hilton, Airtable and the Vancouver Mall are some of the businesses that already use Verkada cameras.

Bottom-up Digital Transformation

 

Next47’s portfolio also contains several startups that use digital tools to open up new potential in industrial manufacturing, whether in additive manufacturing or in specific issues that arise due to the automation of production. Three of them – Markforged, Tulip and Identify3D – presented their technologies at the Hannover Messe in April.

 

“We want to enable digital transformation from the bottom up and give people in production the means to digitally optimize their processes,” says Tulip CEO Natan Linder. Tulip offers apps with multi-industry applications that integrate and visualize production processes. “On the factory floor, the reality is that a great deal of paper and other manual tools are still being used for monitoring and controlling processes. There is a huge gap between reality and what is possible,” he says.

 

Identify3D steps in at another position in the manufacturing process. According to Joe Inkenbrandt, CEO and co-founder of Identify3D, manufacturing is increasingly being separated from engineering. Often, what is shipped and transferred is not the finished commodity stored in quantity and later distributed. Instead, the knowledge about how something is produced is digitally transmitted and items are produced as needed (manufacturing on demand) – two topics that are very current for Siemens. “Along the digital manufacturing process, companies are therefore vulnerable in terms of quality assurance, for instance, or protection of intellectual property,” explains Inkenbrandt. His company offers protection for the entire digital thread by creating a sort of digital container where production processes can be stored and securely transmitted.

Prototypes: 50 Times Faster and 20 Times Cheaper

 

Markforged has used digital methods to open up tremendous opportunities in 3D printing. The company now has major customers worldwide and promises that its technology can make it possible to produce metal prototypes 50 times faster and 20 times cheaper than with conventional methods. Prototypes are one benefit, final products the other. According to Andrew de Geofroy, vice-president of application engineering at Markforged, the company's printers can now make parts for less than $10 that previously took more than $30,000 to produce. “This has far-reaching implications for our target industries, from automotive and aerospace to healthcare and energy,” says Lak Ananth, CEO and Managing Partner at Next47. “We see customers embedding Markforged into their product development and production processes, tremendously improving speed to market and addressing new opportunities in their industries.”

2018-06-29

Sandra Zistl

Picture credits: from above: 1. RioPatuca Images/Fotolia; 3. Verkada; 4. Verkada; 5. Markforged

 

Subscribe to our Newsletter

Stay up to date at all times: everything you need to know about electrification, automation, and digitalization.