At Costanoa, our core mission is to help entrepreneurs create and transform industries. But when a founder walked in and told us, “We are building products using similar technology used in self-driving cars to fundamentally change the $160B fish farming industry,” we were intrigued — but had all kinds of questions. First to hear that fish farming was a $160B industry. But then we wondered, where and how did computer vision and machine learning apply?
We learned that by 2050, it is estimated there will be a 50% protein gap world-wide, and since fish farming is $160B and the fastest growing sector of food production in the world, it can be a main resource to solve this problem. A typical scale salmon farm, for example, produces over five million kilograms of fish per year: producing more fish than beef.
It turns out that aquaculture is expensive and complex with a myriad of variables. Fish feed by itself can be over 50% of the cost to run a farm. Sea lice has likewise already caused $1B of damage in Norway alone. Then there is water temperature, dissolved oxygen shifts, daylight, fish health and diseases, amount of food to feed and when, intermingling with wild fish populations; it is precisely the kind of problem that lends itself well to machine learning.
Much of fish farming equipment is automated, however, there is one operator offsite that watches a live camera feed from inside the pen — monitoring the fish skin, the water temperature, food pellets falling down, and using their best guesses to to optimize how much feed belongs in a pen. Bryton Shang, Founder of Aquabyte, envisioned a farm in which inexpensive off-the-shelf cameras used computer vision to assess algae bloom by analyzing color, to spot lice before it became an outbreak, and to optimize feed by measuring the actual size of the fish versus how much feed doesn’t appear to be eaten. Combined with machine learning algorithms, over time, the software would only get smarter, increasing yields and lowering costs dramatically. It was one of those instances when as soon as the problems and solutions were explained, it made infinite sense as a brilliant use of what new technology has enabled.
This is why we are so excited to announce our investment in Aquabyte.
Bryton impressed us with his deep computer vision background from Princeton and his cross-domain experience with cancer detection and machine learning. With their understanding of the industry and its problems, Bryton and the team convinced us that they could really build the right solution. Several of the world’s largest aquaculture companies were willing to work with Aquabyte in the early days because they’d demonstrated success deploying in their Potrero Hill lab. Rather than reject outsiders and new technology, the aquaculture industry has embraced Aquabyte because of their thoughtful approach to its most important problems.
The ability to transform an industry and improve business for long-lived industries is a feature not a bug for us. Beyond this, we believe in the ability to build substantial value while enhancing food sustainability — for a company to be mission-driven in a way that serves global goals — is a source of power and advantage. We believe Aquabyte can create meaningful and important impact that can revolutionize an industry and help feed the world, which is why we’ve invested in their earliest seed round of funding. We couldn’t be more excited to partnering with Bryton and the Aquabyte team to help build their company.