Everest Labs, one of our newest portfolio companies, is the latest in a series of investments we’ve made in early-stage startups that use AI, robotics and vision to solve industrial sized problems.

The company does for the recycling industry what Canvas does for factories; Osara does for warehouses; and Tortuga does for farms: make operations more efficient and safer. (In fact, we were introduced to Everest by our co-investors with Canvas.)

Robots pluck recyclables from conveyor belt and place them in the proper bin, sorting through waste faster — and, crucially, much safer — than humans. The result: Recycling centers reduce costs, improve productivity and route far more material away from landfills.

Reason 1: It’s a big problem … with a ‘why now’

It turns out, this seemingly simple problem is a particularly costly one for recycling centers, which pay up to $65 an hour to human sorters and collectively lose billions of dollars to errors and inefficiency each year.

The deck summarized the problem with two key stats:

• 34% of non-landfill waste (87 million tons a year) is still processed by humans at material recovery facilities; and  
• $20B+ worth of recyclables end up in landfills every year in the U.S. alone.

There’s also a strong “why now” component to the problem, as China in 2018 upended global trash and recycling markets by banning imports of mixed paper and plastic, and also severely limiting imports of scrap metal. (Waste Management, the largest residential recycler in the U.S., now exports just 3% of its recycled material to China — down from 30% at the peak.)

With fewer options to export trash, recycling centers are incentivized to make use of as much recycled material as they can.

Reason 2: Robots are the ideal solution

Robots are a particularly good option for sorting trash because it’s such a dangerous and dirty job.

When we visited recycling centers, we easily understood the problem, based on the number of empty slots we saw along massive conveyor belts that carry trash for sorting. It’s common for a worker to get poked by a sharp object and take time off. Despite the unusually high hourly rate, it’s also a hard job to fill because of the danger and poor work environment.

The simple, repetitive nature of the tasks also makes it an ideal fit for robots. For humans, these cause overuse injuries over time that hurt productivity and add to healthcare costs.

There’s also the question of speed. Operators of recycling centers told us their workers pick an average of 40 to 60 items per hour. If a robot arm could do better than that, 24 hours a day, seven days a week without a huge upfront cost, they told us, it would be an easy decision.

Reason 3: RaaS model

The first time we met founders JD Ambati and Pratik Khadloya, they gave us demonstration of an early version of their robot. We immediately got the appeal. Rather than a $300k machine that a recycler would be forced to retrofit its operations around, this was a simple robotic arm that was then paired with advanced sensors and AI.

This scaled-down approach to hardware meshes with what we’ve learned with other robotics investments. We’ve come to understand how important it is that robotic solutions drop easily within a customer’s existing footprint.

This alleviates two of the largest hurdles to adoption: First, the up-front operational cost. Second, ease of deployment. Most companies in this space require an enormous up-front cost — the cost of the machine and retrofitting — in exchange for promised savings down the road. Everest offers a RaaS (robots as a service) model: a monthly operational fee with no up-front capital costs.

Reason 4: The right team

Ambati and Khadloya had the combination of business and technical expertise that we look for in a founding team. 

Ambati had executed go-to-market strategies for startups in prior sales and business development roles — including several exits and at least one IPO. Khadloya was steeped in AI, robotics and optics, and he demonstrated that he understood the complexities of offering robotics as a service.

Also, they had worked together at previous companies, we felt they could deliver.

Reason 5: They have interested customers 

During our due diligence, we spoke with several of the largest recycling companies in the U.S. They confirmed the challenges they’re confronting and talked us through what they liked and didn’t like about the many competitor options out there. Two of them offered up their facilities for testing the Everest Labs robots.

These were terrific signs that Everest Labs was onto something. 

Reason 6: Beyond recycling

Waste management isn’t an area we’d typically gravitate to. But this investment fits nicely with other investments we’ve made combining AI, robotics and optics. We also saw the potential for Everest Labs to expand into other verticals — anywhere you see a large conveyor with people lining its sides.