Why We’re Backing Altitude
The intelligence revolution, initially marked by the automation of rote tasks, and more recently indicated by the maturation of large language models, is starting to show what’s possible in the disruption of new industry categories. Robots are already sorting our drugs at pharmaceutical plants, painting our cars inside factories, and automating our 3PL warehouses. Yet there are plenty of jobs that robots, or rather, computer vision and robotics experts, have yet to replicate. Tasks that require a robot to handle objects of varying sizes and shapes with exceptional dexterity have been an unceasing headache. In turn, many of the most unpleasant jobs humans work remain underserved by automation, with robotic arms busy picking, packing, and not doing much else in the industries most desperate for their aid.
The Amplify team is beyond thrilled to be backing Altitude AI in its mission to accelerate the fourth industrial revolution. Altitude’s proprietary SaaS synthesizes advanced computer vision and sensor data to enable off-the-shelf robotic hardware to perceive and manipulate objects like never before. In the long-term, the Altitude operating system will develop and deploy industrial robotic applications across many industries. But to start, they’re zeroing in on where the pain is felt most acutely, in food processing.
The food supply chain as a whole is champing at the bit for automation. Produce growers are hoping robotics will touch everything from planting to harvesting to weeding. Protein processors are keen to expand beyond pick and place. Across verticals, food processors are on break-neck deadlines to implement robotics throughout their operations, with many appointing heads of automation. Yet until very recently, robotics hadn’t been able to consistently move beyond pick and place and warehouse work.
As their initial go to market, Altitude is zeroing in on meat processing. For a sense of scale, the meat industry alone employs more workers than automotive manufacturing in this country. Thus far, nothing has been sophisticated enough to step in along the line itself. In part, this was because of the peculiar challenges posed by the degrees of variation in food products. Depending on the protein, the same cut can vary dramatically from animal to animal (with lamb and fish being among the worst). This degree of variability has long been a barrier to entry for robotic trimming and slicing as the variable shapes and sizes of meat require 3D computer vision for manipulation.
Yet, from conversations, it was clear that the greatest problem facing the industry was staffing and labor along the processing line. Companies in the space revealed shocking rates of worker churn (clocking in at an industry average of 160%), with multiple executives speculating that nobody is operating their plants at above 80% capacity. Churn is only exacerbated by the dangerous, dirty, and shoulder to shoulder work environment typical across the industry.
Diminished capacity is only compounded by an increasingly unskilled workforce. Older, more experienced workers are retiring, and their expertise is leaving with them. If meat processing wasn’t entirely a pennies and ounces business, then workers would be largely interchangeable. But yield is everything, and unskilled workers often cut too much protein in the process of trimming fat. In a revelatory exchange, a CEO of a pork processing plant pointed to a station wherein a flank had been cut too deeply and reckoned that over the course of a year, the cumulative loss of yield from that station alone would amount to $1M in lost revenue for the company. Decreased capacity, worker churn, and dropping yield are all cutting into the already razor thin margins of protein processors.
There is a multi-tiered benefit to implementing robotics along the line itself. Labor turnover is only getting worse and labor costs are climbing. Yield optimization presents a huge opportunity for cost savings as most plants are aware that their workers cut away plenty of good product. Then there’s the food and human safety issue. For food safety reasons, the industry already tries to minimize contact points between workers and product as much as possible. From a worker safety perspective, the facilities are very dangerous, with business owners eager to boost safety. Most trimming and cutting lines have people shoulder-to-shoulder with sharp instruments. The plants are kept at frigid temperatures, and the work can get dirty. It’s not an easy job, and it’s not a particularly safe one.
All of the above is what gets us so excited to be backing Ed Walker as he and the team build Altitude. Ed has spent his entire career focusing on computer perception system design, with a particular interest in multi-sensor modalities. Graduating from Princeton for Computer Science, he joined Google as a software engineer. Later, he spent time at Waymo and Built Robotics, but was always passionate about his own project ideas and always had things percolating in spare time. Ed’s original idea was to build a robotics system to create fast casual food experiences, using existing hardware and novel software. After exploring in industry, he realized that labor was important for the fast casual space, but was spread across many use cases so it was hard to find one that took a significant enough portion of COGS off the table. An advisor connected him with a meatpacker, and it was a lightbulb moment. Nearly all of the cost outside of raw materials is the manual-intensive labor around things like trimming and cutting. Ed thought he could adapt a lot of the work and research he was already doing to fit the space. From there, Altitude as we know it was off to the races.
Proprietary sensor data is ingested, the Altitude software makes sense of the data with in-house perception algorithms, and movement commands informed by the results of the prior step are sent to a robotic arm. Altitude is not the eyes, nor the arms of industrial automation, it’s the brains!
Industry appetite, coupled with Ed’s technical prowess, is an exciting combination. Implementing robotics software that limits food and worker safety hazards, increases yield, and supplements labor with human level dexterity, is what Altitude AI is already delivering in a series of pilots. This new infusion of cash will allow the Salt Lake City based team to expand existing contracts and meet preliminary demand.
The team recently closed on $3.165M from Amplify and co-investors we’re lucky to work with: Kickstart Fund, Lerer Hippeau, and Signal Peak Ventures.
If you found this valuable, consider giving this a share, and passing it along to anyone who you think may find this relevant.
Amplify is a first check fund backing LA’s earliest and strongest tech teams. You can get in touch with us here. Want these updates delivered to your inbox? Join our newsletter!
