Revise Robotics Wants Robots to Refurbish Your Old Laptop So Humans Don't Have To

The Macro: E-Waste Is a $50 Billion Problem Hiding in Plain Sight

Every year, more than 60 million tons of electronics get discarded globally. Most of that ends up in landfills, shipped to developing countries for unsafe manual processing, or sitting in a drawer in your house because you’re not sure what to do with your 2019 MacBook Pro. The recoverable value in those devices is staggering. We’re talking about $50 billion worth of usable components, screens, batteries, and intact machines that could be refurbished and resold.

The refurbishment industry exists, but it’s almost entirely manual. A human picks up a laptop. A human opens it, runs diagnostics, wipes the drive, tests the keyboard and trackpad, checks the screen for dead pixels, photographs it from six angles, writes a description, and lists it for resale. Each unit takes 30 to 60 minutes of skilled labor. The economics work when you’re paying $15 an hour in the US, but they’re thin, and the throughput ceiling is low.

Companies like Back Market, Gazelle, and Decluttr have built real businesses on the demand side of refurbished electronics. The consumer appetite is there. Refurbished laptops and phones sell well because people like saving money and feeling mildly good about the environment. The bottleneck has always been supply-side processing. You can’t refurbish more units than you have hands to touch them.

This is exactly the kind of problem where robotics should work. The tasks are repetitive, the physical manipulation requirements are moderate (you’re not assembling a car, you’re opening a laptop), and the quality control can be handled by computer vision. The question has always been whether the hardware and AI are good enough to handle the variation. Laptops come in hundreds of form factors. They arrive in unknown condition. Some are pristine, some have cracked screens and missing keys.

The Micro: Robots That Do the Whole Job

Revise Robotics is building an AI-enabled robotic system designed to sit inside e-waste facilities and process mixed batches of laptops from start to finish. The system tests hardware, wipes drives, photographs devices, grades condition, and posts them for resale. The target is hundreds of units per day with zero human intervention.

That last part is the ambitious bit. Not “less human intervention” or “human-assisted robotics.” Zero. The robot picks up a laptop it’s never seen before, figures out what it is, determines if it works, and either prepares it for resale or routes it for parts harvesting.

Rupesh Jeyaram and Antonio Monreal are the founders. Rupesh has a Wharton MBA (2024) and a Caltech degree (2020). Antonio is CTO. They’re a three-person team that came through YC’s Winter 2025 batch with Garry Tan as their partner. Rupesh’s background straddles business and engineering, which makes sense for a company that needs to sell into industrial facilities while also solving hard robotics problems.

The Caltech connection is worth noting because Caltech has one of the strongest robotics and computer vision programs in the country. This isn’t a software team trying to bolt a robot arm onto a conveyor belt. They’re building from genuine technical depth.

The competitive space for electronics refurbishment automation is surprisingly thin. Most refurbishment companies are labor operations with some software tools. ERPA (Electronic Refurbishment and Processing Association) members are mostly manual operations. Li-Cycle and Redwood Materials do battery recycling with some automation, but they’re focused on raw materials recovery, not device-level refurbishment. The idea of a fully automated refurbishment cell for consumer electronics is genuinely new.

The $1 trillion number in their pitch refers to the cumulative value of discarded electronics, not annual revenue potential, but even a small slice of that is a big market. Refurbished electronics is a $150 billion global market growing at 11-12% annually. If Revise can meaningfully increase throughput at refurbishment facilities, they’re capturing value on both sides: more units processed per facility, and higher margins because labor cost drops to nearly zero per unit.

The Verdict

This is one of those companies where the ambition and the market opportunity are clearly real, and the question is purely about technical execution. Can the robots actually handle the variation? A ThinkPad from 2018 and a MacBook Air from 2022 are physically different in every way. Can the system reliably wipe an encrypted drive when it doesn’t know the OS? Can the photography and grading be consistent enough that buyers trust the listings?

I think the technical risk is real but bounded. Computer vision is good enough to identify and grade consumer electronics. Robotic manipulation for laptop-scale objects is within the current state of the art. The hard part is doing all of it reliably, at speed, in a factory environment where laptops arrive in random condition covered in someone’s stickers.

At 30 days, I’d want to know their unit processing time and first-pass yield rate. How many laptops does the system handle per hour, and what percentage need human intervention? At 60 days, I’d want to see resale prices compared to human-refurbished equivalents. If the grading is accurate, prices should be comparable. At 90 days, the question is whether e-waste facilities are willing to install the system and trust it with real throughput. The economics should be compelling on paper. Whether they’re compelling enough to overcome industrial risk-aversion is the real test.