fbpx

AMP Robotics Amped Over AI-Powered Recycling Robots

We recently wrote about the nascent and unprofitable carbon capture market – and suddenly it seemed like everyone was talking about how we need new, scalable technologies for removing and recycling CO2 into everything from energy to cement. It didn’t hurt that Mr. Elon Musk happened to announce the same day that he was putting up $100 million in prize money for the best carbon capture technologies to help spur the industry. It’s an interesting move from the eccentric billionaire, whose electric vehicle company, Tesla (TSLA), earns hundreds of millions of dollars by selling carbon credits to other automakers, so Musk has good reason to invest in green technology as head of the most valuable company (at least as of today) in the sector.

There are plenty of other reasons to believe that the green economy is about to bloom, such as the 10 million new clean energy jobs the latest guy in the White House just promised. Or the $55 million a Denver startup just raised for AI-powered recycling robots.

Recycling and the Green Economy

Usually, in these kinds of articles, we’d hit you with a few U.S. statistics like the current recycling rate (about 32%) and the rough number of people employed (about 680,000 jobs, representing $37.8 billion in wages) in the industry who are soon to be replaced by robots. We’d also rehash the big shake-up to the global recycling industry when China stopped being the world’s depository for poorly sorted recyclables starting in 2018. And then we’d  launch into an inspiring story about a company that stands ready to save the Earth using artificial intelligence, advanced robotics, and a significant amount of someone else’s money in the pursuit of market share. But before we talk about a pretty interesting company from our favorite Cow Town, we should quickly consider a fundamental question: What is garbage worth? 

Chart showing types of waste streams through time.
Recycling and waste-to-energy remain small markets in the garbage business. Credit: EPA

We quickly discovered that this was a really dumb question to ask, because it’s the sort of topic that could send an MBA down a rabbit hole, writing unwieldy reports for McKinsey until sweet, sweet death ended it all. Waste disposal and recycling is a highly fragmented industry with vastly varying degrees of regulations, subsidies, and other factors that make it pretty daunting to understand the current state of the market, let alone its potential. You have everything from waste-to-energy plays and biofuel to new technologies for recycling plastics or replacing them altogether

Another way to approach the problem is to look at the industry that the green economy is trying to replace – fossil fuels. That’s also not an easy number to come by, but one reasonable estimate that was put out there a few years ago was that there are nearly 1,500 oil and gas firms listed on stock exchanges around the world, worth a combined $4.65 trillion. That was way back in 2014 when ExxonMobil (XOM) boasted a market cap of $425 billion. Today, the company is valued closer to $200 billion, as energy slumped during the pandemic and green tech like solar and wind matured and grabbed market share. A recent estimate put total revenues for oil and gas at about $3.3 trillion. 

Chart showing rise in renewable energy market.
Renewable energy is fast catching up with fossil fuels, as part of the growing green economy. Credit: Statista

Of course, there’s a difference between value and revenues – and it’s not an apples-to-apples comparison by any means – but it gives us a rough idea of how much value the green economy must replace. The size of the renewable energy market alone is expected to be worth north of $1.5 trillion by 2025. In addition, the electric vehicle market is starting to show the kind of growth we’d expect from an emerging technology sector that has entered the disruptive phase. Expect the announcement by General Motors (GM) to turn its entire fleet into EVs by 2035 to only accelerate that market into trillion-dollar territory by the end of this decade. So, while no one is predicting that recycling a few more plastic bottles will change the world overnight, it’s not unreasonable to believe that it could become a viable sector within the larger green economy.

An AI-Powered Recycling Robotic Startup

Certainly a key to kickstarting a more robust recycling market is to ensure cheap raw materials for end-user companies. Of course, people can’t be trusted with the simplest tasks, like sorting their trash correctly, so most recyclables end up in the landfill anyway. Technologies like recycling vending machines can provide a steady source of some raw materials, but end-users have to incentivize consumers to participate and pay a fee for the privilege. The better solution is to pluck out the choicest raw materials from a single-stream waste source after it reaches a waste-sorting facility. That requires some fast and smart robots.

Click for company website

Founded in 2015, AMP Robotics has raised $74.5 million in disclosed funding, including $55 million from a Series B last month that included high-profile venture capital firm Sequoia Capital and Google’s VC arm. We briefly profiled AMP Robotics exactly three years ago in an article about AI-powered waste sorting robots. The other two companies we covered – Sadako Technologies and ZenRobotics – continue to compete in the same space, but neither seems to have the same level of financing or high-profile clients as AMP Robotics. 

An AI Platform for Identifying Recyclables

Founder and CEO Matanya Horowitz developed the technology that would eventually power his company during his time at Caltech, and later while competing in robotics challenges sponsored by the shadowy government agency known as DARPA. For the latter, he worked on projects to build better robotic arms, as well as machines that could perform human-like tasks in dangerous situations like disabling bombs or breaking up with a girlfriend.  

Computer vision for recycling.
Computer vision for recycling. Credit: AMP Robotics

AMP Robotics calls the result of all that R&D Neuron, which “applies computer vision and deep learning to guide high-speed robotics systems to precisely identify and differentiate recyclables found in the waste stream by color, size, shape, opacity, consumer brand, and more, storing data about each item it perceives.” It can snatch stuff as small as a bottle cap off of a conveyor belt at a pick rate of upwards of 80 items per minute, which is more than twice as fast as forced child laborers human sorters. The company calls the high-speed robotics system itself AMP Cortex:

AMP Robotics AI-powered robotic recycling system.
No, it’s not a Jewish space laser, just the latest AI-powered robotic recycling system from AMP Robotics. Credit: AMP Robotics

The AI platform Neuron uses machine learning to improve over time, integrating data from its Cortex robotic systems deployed to different customers around the world. The AI can precisely identify different types of plastics, including polyethylene terephthalate (PET), high-density polyethylene (HDPE), low-density polyethylene (LDPE), polypropylene, and polystyrene.  It can even sort these plastics by color, clarity, and opacity, along with different form factors, such as lids versus tubs versus clamshells and cups.

Building the Business

Like many companies in the Age of Rona, AMP Robotics has benefitted from wider adoption of automation technologies. The startup was even featured in an article by the New York Times, which quoted Horowitz that his company was seeing a “significant” bump in orders for its Cortex robotic systems during the early days of the pandemic. Apparently demand never slowed. Late in 2020, the company inked a long-term deal with Waste Connections (WCN), a $25 billion waste management company that operates throughout the United States and Canada, for 24 AI-guided robotics systems, the company’s largest contract to date.

AMP Robotics AI-guided waste sorting robots.
Credit: AMP Robotics

The other big news last year was the announcement of one of AMP’s first corporate partners – consumer beverage giant Keurig Dr Pepper (KDP). The AI-powered recycling robots are now trained to properly identify and sort K-Cup coffee pods in recycling facilities. It’s a small item with a huge environmental impact. Before KDP made the K-Cup pods full recyclable, billions were sent to landfills – enough to circle the Earth 10 times in a decade of caffeine consumption.

AMP’s robots are deployed globally, including in more than 20 U.S. states, as well as Asia and Europe. It recently installed a facility in Spain and also opened a 40,000-square-foot test facility in Colorado last year. And the startup is expanding its AI capabilities for construction and demolition debris

As one happy customer reported:

We’re running at full capacity because AMP’s robots are there all day long, and the robots aren’t charging us overtime. They work at the same speed consistently, whether fresh in the morning from a cup of coffee or at the end of a 10-hour shift. Without the robots, we’d see residue coming off the line and items that should’ve been caught. We don’t see that anymore. AMP’s robots are designed to operate in tough conditions, like the harsh environments you come across in recycling.

A Future Robot Overlord

Conclusion

AMP Robotics claims recycling could be big business, tossing out an unsubstantiated number of $200 billion. We’d certainly like to learn more about the assumptions and timeline behind that figure, but we’re pretty sure it won’t happen without significant government subsidies. While all of you free-market types might guffaw at the notion, keep in mind that even the most mature markets still take plenty of government hand-outs. Remember all the numbers we threw at you earlier about the oil and gas industry, well here’s one more: The International Monetary Fund estimates that fossil fuels got about $5.2 trillion in subsidies in 2017 alone. One industry’s welfare system is another industry’s economic incentives.

Share

Leave a Reply

Your email address will not be published.