Industrial drone maker Percepto raises $45M and integrates with Boston Dynamics’ Spot

Consumer drones have over the years struggled with an image of being no more than expensive and delicate toys. But applications in industrial, military and enterprise scenarios have shown that there is indeed a market for unmanned aerial vehicles, and today, a startup that makes drones for some of those latter purposes is announcing a large round of funding and a partnership that provides a picture of how the drone industry will look in years to come.

Percepto, which makes drones — both the hardware and software — to monitor and analyze industrial sites and other physical work areas largely unattended by people, has raised $45 million in a Series B round of funding.

Alongside this, it is now working with Boston Dynamics  and has integrated its Spot robots with Percepto’s Sparrow drones, with the aim being better infrastructure assessments, and potentially more as Spot’s agility improves.

The funding is being led by a strategic backer, Koch Disruptive Technologies, the investment arm of industrial giant Koch Industries (which has interests in energy, minerals, chemicals and related areas), with participation also from new investors State of Mind Ventures, Atento Capital, Summit Peak Investments, Delek-US. Previous investors U.S. Venture Partners, Spider Capital and Arkin Holdings also participated. (It appears that Boston Dynamics and SoftBank are not part of this investment.)

Israel-based Percepto has now raised $72.5 million since it was founded in 2014, and it’s not disclosing its valuation, but CEO and founder Dor Abuhasira described as “a very good round.”

“It gives us the ability to create a category leader,” Abuhasira said in an interview. It has customers in around 10 countries, with the list including ENEL, Florida Power and Light and Verizon.

While some drone makers have focused on building hardware, and others are working specifically on the analytics, computer vision and other critical technology that needs to be in place on the software side for drones to work correctly and safely, Percepto has taken what I referred to, and Abuhasira confirmed, as the “Apple approach”: vertical integration as far as Percepto can take it on its own.

That has included hiring teams with specializations in AI, computer vision, navigation and analytics as well as those strong in industrial hardware — all strong areas in the Israel tech landscape, by virtue of it being so closely tied with its military investments. (Note: Percepto does not make its own chips: these are currently acquired from Nvidia, he confirmed to me.)

“The Apple approach is the only one that works in drones,” he said. “That’s because it is all still too complicated. For those offering an Android-style approach, there are cracks in the complete flow.”

It presents the product as a “drone-in-a-box”, which means in part that those buying it have little work to do to set it up to work, but also refers to how it works: its drones leave the box to make a flight to collect data, and then return to the box to recharge and transfer more information, alongside the data that is picked up in real time.

The drones themselves operate on an on-demand basis: they fly in part for regular monitoring, to detect changes that could point to issues; and they can also be launched to collect data as a result of engineers requesting information. The product is marketed by Percepto as “AIM”, short for autonomous site inspection and monitoring.

News broke last week that Amazon has been reorganising its Prime Air efforts — one sign of how some more consumer-facing business applications — despite many developments — may still have some turbulence ahead before they are commercially viable. Businesses like Percepto’s stand in contrast to that, with their focus specifically on flying over, and collecting data, in areas where there are precisely no people present.

It has dovetailed with a bigger focus from industries on the efficiencies (and cost savings) you can get with automation, which in turn has become the centerpiece of how industry is investing in the buzz phrase of the moment, “digital transformation.”

“We believe Percepto AIM addresses a multi-billion-dollar issue for numerous industries and will change the way manufacturing sites are managed in the IoT, Industry 4.0 era,” said Chase Koch, President of Koch Disruptive Technologies, in a statement. “Percepto’s track record in autonomous technology and data analytics is impressive, and we believe it is uniquely positioned to deliver the remote operations center of the future. We look forward to partnering with the Percepto team to make this happen.”

The partnership with Boston Dynamics is notable for a couple of reasons: it speaks to how various robotics hardware will work together in tandem in an automated, unmanned world; and it speaks to how Boston Dynamics is pulling up its socks.

On the latter front, the company has been making waves in the world of robotics for years, specifically with its agile and strong dog-like (with names like “Spot” and “Big Dog”) robots that can cover rugged terrains and handle tussles without falling apart.

That led it into the arms of Google, which acquired it as part of its own secretive moonshot efforts, in 2013. That never panned out into a business, and probably gave Google more complicated optics at a time when it was already being seen as too powerful. Then, SoftBank stepped in to pick it up, along with other robotics assets, in 2017. That hasn’t really gone anywhere either, it seems, and just this month it was reported that Boston Dynamics was reportedly facing yet another suitor, Hyundai.

All of this is to say that partnerships with third parties that are going places (quite literally) become strong signs of how Boston Dynamics’ extensive R&D investments might finally pay off with enterprising dividends.

Indeed, while Percepto has focused on its own vertical integration, longer term and more generally there is an argument to be made for more interoperability and collaboration between the various companies building “connected” and smart hardware for industrial, physical applications. It means that specific industries can focus on the special equipment and expertise they require, while at the same time complementing that with hardware and software that are recognised as best-in-class. Abuhasira said that he expects the Boston Dynamics partnership to be the first of many.

That makes this first one an interesting template. It will see Spot carrying Percepto’s payloads for high resolution imaging and thermal vision “to detect issues including hot spots on machines or electrical conductors, water and steam leaks around plants and equipment with degraded performance, with the data relayed via AIM.” It will also mean a more thorough picture, beyond what you get from the air, and potentially a point at which the data that the pairing sources results even in repairs or other work to fix issues.

“Combining Percepto’s Sparrow drone with Spot creates a unique solution for remote inspection,” said Michael Perry, VP of Business Development at Boston Dynamics, in a statement. “This partnership demonstrates the value of harnessing robotic collaborations and the insurmountable benefits to worker safety and cost savings that robotics can bring to industries that involve hazardous or remote work.”

Recycling robotics company AMP Robotics could raise up to $70M

AMP Robotics, the recycling robotics technology developer backed by investors including Sequoia Capital and Sidewalk Infrastructure Partners, is close to closing on as much as $70 million in new financing, according to multiple sources with knowledge of the company’s plans.

The new financing speaks to AMP Robotics’ continued success in pilot projects and with new partnerships that are exponentially expanding the company’s deployments.

Earlier this month the company announced a new deal that represented its largest purchase order for its trash sorting and recycling robots.

That order, for 24 machine learning-enabled robotic recycling systems with the waste handling company Waste Connections, was a showcase for the efficacy of the company’s recycling technology.

That comes on the back of a pilot program earlier in the year with one Toronto apartment complex, where the complex’s tenants were able to opt into a program that would share recycling habits monitored by AMP Robotics with the building’s renters in an effort to improve their recycling behavior.

The potential benefits of AMP Robotic’s machine learning enabled robots are undeniable. The company’s technology can sort waste streams in ways that traditional systems never could and at a cost that’s far lower than most waste handling facilities.

As TechCrunch reported earlier the tech can tell the difference between high-density polyethylene and polyethylene terephthalate, low-density polyethylene, polypropylene and polystyrene. The robots can also sort for color, clarity, opacity and shapes like lids, tubs, clamshells and cups — the robots can even identify the brands on packaging.

AMP’s robots already have been deployed in North America, Asia and Europe, with recent installations in Spain and across the U.S. in California, Colorado, Florida, Minnesota, Michigan, New York, Texas, Virginia and Wisconsin.

At the beginning of the year, AMP Robotics  worked with its investor, Sidewalk Labs on a pilot program that provided residents of a single apartment building representing 250 units in Toronto with detailed information about their recycling habits. Sidewalk Labs is transporting the waste to a Canada Fibers material recovery facility where trash is sorted by both Canada Fibers employees and AMP Robotics.

Once the waste is categorized, sorted and recorded, Sidewalk communicates with residents of the building about how they’re doing in their recycling efforts.

It was only last November that the Denver-based AMP Robotics raised a $16 million round from Sequoia Capital and others to finance the early commercialization of its technology.


As TechCrunch reported at the time, recycling businesses used to be able to rely on China to buy up any waste stream (no matter the quality of the material). However, about two years ago, China decided it would no longer serve as the world’s garbage dump and put strict standards in place for the kinds of raw materials it would be willing to receive from other countries.

The result has been higher costs at recycling facilities, which actually are now required to sort their garbage more effectively. At the time, unemployment rates put the squeeze on labor availability at facilities where trash was sorted. Over the past year, the COVID-19 pandemic has put even more pressure on those recycling and waste handling facilities, despite their identification as “essential workers”.

Given the economic reality, recyclers are turning to AMP’s technology — a combination of computer vision, machine learning and robotic automation to improve efficiencies at their facilities.

And, the power of AMP’s technology to identify waste products in a stream has other benefits, according to chief executive Matanya Horowitz.

“We can identify… whether it’s a Coke or Pepsi can or a Starbucks cup,” Horowitz told TechCrunch last year. “So that people can help design their product for circularity… we’re building out our reporting capabilities and that, to them, is something that is of high interest.”

AMP Robotics declined to comment for this article.


Relativity Space raises $500 million as its sets sights on the industrialization of Mars

3D-printed rocket startup Relativity Space has closed $500 million in Series D funding (making official the earlier reported raise), the company announced today. This funding was led by Tiger Global Management, and included participation by a host of new investors including Fidelity Management & Research Company, Baillie Gifford, Iconiq Capital, General Catalist and more. This brings the company’s total raised so far to nearly $700 million, as the startup is poised to launch its first ever fully 3D-printed orbital rocket next year.

LA-based Relativity had a big 2020, completing work on a new 120,000 square-foot manufacturing facility in Long Beach. Its rocket construction technology, which is grounded in its development and use of the largest metal 3D printers in existence, suffered relatively few setbacks due to COVID-19-related shutdowns and work stoppages since it involves relatively few actual people on the factory floor managing the 3D printing process, which is handled in large part by autonomous robotic systems and software developed by the company.

Relativity also locked in a first official contract from the U.S. government this year, to launch a new experimental cryogenic fluid management system on behalf of client Lockheed Martin, as part of NASA’s suite of Tipping Point contracts to fund the development of new technologies for space exploration. It also put into service its third-generation Stargate 3D metal printers – the largest on Earth, as mentioned.

The company’s ambitions are big, so this new large funding round should provide it with fuel to grow even more aggressively in 2021. It’s got new planned initiatives underway, both terrestrial and space-related, but CEO and founder Tim Ellis specifically referred to Mars and sustainable operations on the red planet as one possible application of Relativity’s tech down the road.

In prior conversations, Ellis has alluded to the potential for Relativity’s printers when applied to other large-scale metal manufacturing – noting that the cost curve as it stands makes most sense for rocketry, but could apply to other industries easily as the technology matures. Whether on Mars or on Earth, large-scale 3D printing definitely has a promising future, and it looks like Relativity is well-positioned to take advantage.

We’ll be talking to Ellis at our forthcoming TC Sessions: Space event, so we’ll ask him more about this round and his company’s aspirations live there, too.

Interlocking AIs let robots pick and place faster than ever

One of the jobs for which robots are best suited is the tedious, repetitive “pick and place” task common in warehouses — but humans are still much better at it. UC Berkeley researchers are picking up the pace with a pair of machine learning models that work together to let a robot arm plan its grasp and path in just milliseconds.

People don’t have to think hard about how to pick up an object and put it down somewhere else — it’s not only something we’ve had years of practice doing every day, but our senses and brains are well adapted for the task. No one thinks, “what if I picked up the cup, then jerked it really far up and then sideways, then really slowly down onto the table” — the paths we might move an object along are limited and usually pretty efficient.

Robots, however, don’t have common sense or intuition. Lacking an “obvious” solution, they need to evaluate thousands of potential paths for picking up an object and moving it, and that involves calculating the forces involved, potential collisions, whether it affects the type of grip that should be used, and so on.

Once the robot decides what to do it can execute quickly, but that decision takes time — several seconds at best, and possibly much more depending on the situation. Fortunately, roboticists at UC Berkeley have come up with a solution that cuts the time needed to do it by about 99 percent.

The system uses two machine learning models working in relay. The first is a rapid-fire generator of potential paths for the robot arm to take based on tons of example movements. It creates a bunch of options, and a second ML model, trained to pick the best, chooses from among them. This path tends to be a bit rough, however, and needs fine-tuning by a dedicated motion planner — but since the motion planner is given a “warm start” with the general shape of the path that needs to be taken, its finishing touch is only a moment’s work.

Diagram showing the decision process – the first agent creates potential paths and the second selects the best. A third system optimizes the selected path.

If the motion planner was working on its own, it tended to take between 10 and 40 seconds to finish. With the warm start, however, it rarely took more than a tenth of a second.

That’s a benchtop calculation, however, and not what you’d see in an actual warehouse floor situation. The robot in the real world also has to actually accomplish the task, which can only be done so fast. But even if the motion planning period in a real world environment was only two or three seconds, reducing that to near zero adds up extremely fast.

“Every second counts. Current systems spend up to half their cycle time on motion planning, so this method has potential to dramatically speed up picks per hour,” said lab director and senior author Ken Goldberg. Sensing the environment properly is also time-consuming but being sped up by improved computer vision capabilities, he added.

Right now robots doing pick and place are nowhere near the efficiency of humans, but small improvements will combine to make them competitive and, eventually, more than competitive. The work when done by humans is dangerous and tiring, yet millions do it worldwide because there’s no other way to fill the demand created by the growing online retail economy.

The team’s research is published this week in the journal Science Robotics.

Astroscale sets March 2021 for first commercial orbital debris removal demonstration

Japanese startup Astroscale is aiming for March 2021 for a launch of its first-ever active orbital debris removal mission. This demonstration of its technology, which it hopes to use to help ensure that low-Earth orbit becomes a sustainable environment for commercial activity as it becomes increasingly crowded thanks to the rapid pace of new spacecraft launches.

This demonstration mission, which is called the “End-of-Life Services by Astroscale-demonstration” (ELSA-d for short) will take off from Kazakhstan, launched via a Russian Soyuz rocket. The actual demonstration itself will see Astroscale’s payload, which includes both a ‘servicer’ (which represents the actual debris removal component) and a ‘client’ (which represents any potential satellite or space junk that Astroscale might eventually be tasked with removing).

The servicer unit will use magnets to ‘capture’ the client, docking with it multiple times to show its efficacy, while the client remains stationary and while it emulates an end-over-end tumbling motion that is common for a lot of defunct orbital debris. The purpose of the mission is to show that Astroscale’s technology for seeking out and finding targets for removal, as well as proper target identity verification and docking/release procedures all work as the startup intended.

Low-Earth orbit space junk removal is half of Astroscale’s approach to making space more sustainable for commercial and research activities – the other is on-orbit servicing of geostationary satellites, which tend to be larger and more expensive and occupy an orbital band deeper out in space. The company recently acquired assets of an Israeli company focused on that endeavor in order to bolster that parallel mission.

Sequoia-backed recycling robot maker AMP Robotics gets its largest purchase order

AMP Robotics, the manufacturer of robotic recycling systems, has received its largest purchase order from the publicly traded North American waste handling company, Waste Connections.

The order, for 24 machine learning enabled robotic recycling systems, will be used on container, fiber and residue lines across numerous materials recovery facilities, the company said.

The AMP technology can be used to recover plastics, cardboard, paper, cans, cartons and many other containers and packaging types reclaimed for raw material processing.

The tech can tell the difference between high-density polyethylene and polyethylene terephthalate, low-density polyethylene, polypropylene, and polystyrene. The robots can also sort for color, clarity, opacity and shapes like lids, tubs, clamshells, and cups — the robots can even identify the brands on packaging.

So far, AMP’s robots have been deployed in North America, Asia, and Europe with recent installations in Spain, and across the US in California, Colorado, Florida, Minnesota, Michigan, New York, Texas, Virginia and Wisconsin.

In January, before the pandemic began, AMP Robotics worked with its investor, Sidewalk Labs on a pilot program that would provide residents of a single apartment building representing 250 units in Toronto with detailed information about their recycling habits.

Working with the building and a waste hauler, Sidewalk Labs  would transport the waste to a Canada Fibers material recovery facility where trash will be sorted by both Canada Fibers employees and AMP Robotics. Once the waste is categorized, sorted, and recorded Sidewalk will communicate with residents of the building about how they’re doing in their recycling efforts.

Sidewalk says that the tips will be communicated through email, an online portal, and signage throughout the building every two weeks over a three-month period.

For residents, it was an opportunity to have a better handle on what they can and can’t recycle and Sidewalk Labs is betting that the information will help residents improve their habits. And for folks who don’t want their trash to be monitored and sorted, they could opt out of the program.

Recyclers like Waste Connections should welcome the commercialization of robots tackling industry problems. Their once-stable business has been turned on its head by trade wars and low unemployment. About two years ago, China decided it would no longer serve as the world’s garbage dump and put strict standards in place for the kinds of raw materials it would be willing to receive from other countries. The result has been higher costs at recycling facilities, which actually are now required to sort their garbage more effectively.

At the same time, low unemployment rates are putting the squeeze on labor availability at facilities where humans are basically required to hand-sort garbage into recyclable materials and trash.

AMP Robotics is backed by Sequoia Capital,  BV, Closed Loop Partners, Congruent Ventures  and Sidewalk Infrastructure Partners, a spin-out from Alphabet that invests in technologies and new infrastructure projects.

Pioneers of in-space refueling and manufacturing join TC Sessions: Space 2020

One of the problems with putting a satellite in orbit is that once you do, it’s pretty much out of your hands. If anything goes wrong, or it runs out of fuel, that’s all she wrote. Fortunately there are companies that aim to change this, and three leaders in the field — Orbit Fab, Astroscale, and Maxar — will be joining us at TC Sessions: Space in December.

Also at TC Sessions: Space

You may remember Orbit Fab from Disrupt’s Startup Battlefield around this time last year. CEO and co-founder Daniel Faber debuted its refueling interface, RAFTI, and showed how that and a network of “tanker” satellites could save companies hundreds of millions by keeping their spacecraft in orbit rather than sending up replacements.

Astroscale is embarking on a similar effort for satellites in geosynchronous orbits, which are even more expensive to replace. But the Japan-based company is it’s also aiming at taking down the innumerable dead satellites and debris scattered throughout other orbits, and has raised huge sums to do so. US President Ron Lopez will join the panel to discuss the many potential approaches to improving sustainability in space.

Maxar is of course a well-known name in space operations, and we’ve had head of space robotics Lucy Condrakchian on stage at TC Sessions: Robotics. Her team is currently working on the ambitious Restore-L mission, which will demonstrate on-orbit refueling, manufacturing, and assembly. Why build it down here if you can do it up there?

These three panelists will discuss the possibilities of this emerging industry and what it could mean for startups and established enterprises here on the ground. With costs of launch dropping, the cost of building and maintaining a major satellite becomes a greater issue — but tiny, cheap satellites are also beginning to proliferate.

How will the market evolve? Can proprietary but practical tech like RAFTI make a difference? How close are we to the first satellite built entirely in space? All this and more will be on the table for our panel next month.

Get an early-bird ticket for just $125 until this Friday, November 13. And we have discounts available for groupsstudentsactive military/government employees and for early-stage space startup founders who want to pitch and give their startup some extra visibility.



NASA partners with SpaceX, Rocket Lab, Blue Origin and others for test flights and research

NASA has announced 20 new partnerships with commercial space outfits, among them collaborations with SpaceX, Blue Origin, and Rocket Lab. While no money will change hands, NASA will dedicate millions in personnel and other support to these test launches and developing technologies.

The partnerships are NASA’s Announcement of Collaboration Opportunity selections for 2020. These agreements are unlike the SBIR or NIAC programs in that NASA doesn’t just send some money out and say “let us know how you’re getting on in 6 months or so.”

Instead, the space agency offers open access to its facilities and experts, some of which are the most advanced in the world. It’s a true public-private partnership, for which reason it is still a competitive process to get a project approved — and the list of 17 companies includes several large ones.

SpaceX will be working with Langley to monitor and perform thermal measurements of its Starship launch vehicle and spacecraft during reentry operations over the Pacific.

Rocket Lab, similarly, will partner with Langley, Ames, and Armstrong to do analysis of its Electron launch vehicle as it migrates the hardware towards reusability. The company recently moved up the date it would attempt a full booster recovery to just a week from now, but it’s unclear whether this is an operation NASA will be involved in.

Blue Origin, meanwhile, has two separate partnerships. One is another multi-center effort in which the company will be helping develop a “space robot operating system.” This sounds grand but is probably more of an integration effort, bringing together multiple open source and NASA-developed frameworks to work together, reducing costs and improving interoperability.

The other is regarding using 3D printing to improve engine designs; perhaps they regret letting Tim Ellis run away and start Relativity Space — he cut his teeth doing just this kind of work at Blue Origin and now it appears the company is going to have to play catch-up.

The rest of the partnerships, from artificial lunar regolith to radio-frequency propulsion, can be read about at this NASA post.

Provizio closes $6.2M seed round for its car safety platform using sensors and AI

Provizio, a combination hardware and software startup with technology to improve car safety, has closed a seed investment round of $6.2million. Investors include Bobby Hambrick (the founder of Autonomous Stuff); the founders of Movidius; the European Innovation Council (EIC); ACT Venture Capital.

The startup has a ‘five-dimensional’ sensory platform that – it says – perceives, predicts and prevents car accidents in real-time and beyond the line-of-sight. Its ‘Accident Prevention Technology Platform’ combines proprietary vision sensors, machine learning, radar and with ultra-long range and foresight capabilities to prevent collisions at high speed and in all weather conditions, says the company. The Provizio team is made up of experts in robotics, AI, and vision and radar sensor development.

Barry Lunn, CEO of Provizio Said: “One point three five road deaths to zero drives everything we do at Provizio. We have put together an incredible team that is growing daily. AI is the future of automotive accident prevention and Provizio 5D radars with AI on-the-edge are the first step towards that goal.”

Also involved in Provizio is also Dr. Scott Thayer and Prof Jeff Mishler formally of Carnegie Mellon robotics, famous for developing early autonomous technologies for Google/ class="crunchbase-link" href="" target="_blank" data-type="organization" data-entity="waymo">Waymo, Argo, Aurora and Uber.

Autonomous drone startup Skydio taps Tesla, Samsara veterans in enterprise push

Autonomous drone company Skydio has hired three executives in product and engineering following its recent $100 million Series C funding round as part of the company’s strategy to expand beyond consumer applications and into the enterprise and public sector markets.

The hires include Roy Goldman, who was the director of software development at Tesla for five years and most recently held a similarly senior position at Carbon. Goldman has been hired to head up Skydio’s product management.

The company also hired Ryan Reading, who previously worked at Samsara where he has been vice president of engineering and more recently general manager of fleet safety, is now head of software engineering. Mike Ross, who led the he led the telematics product group at Samsara, was hired as senior director of product management.

The company said Thursday that the trio will “play critical roles in realizing the company’s vision for the first-of-its-kind integrated enterprise autonomy stack.”

Skydio CEO Adam Bry noted in a blog post that their track record in delivering enterprise products with cloud-connected hardware will be “key” for the company.

Skydio raised $100 million earlier this year to fund its next phase of product development for the enterprise, public sector and defense markets.

Skydio initially focused on consumer drones, launching two since its founding in 2014. Both models of consumer drones use artificial intelligence technology to fly without a human operator. The autonomous system is able to track objects and people, while simultaneously avoiding potential collisions with objects, including trees, power lines and other obstacles.

The company announced this summer a new X2 drone platform designed for enterprise use. Skydio previously said that the X2 drone, which includes onboard 360-degree superzoom camera, a FLIR 320×256 resolution thermal imaging camera, a battery life of 35 minutes of flying time and a maximum range of 6.2 miles, will ship in fourth quarter of this year.