Deep tech doesn’t have to be capital & time intensive

“If we make mistakes in our first compositions and do not know them, we may not amend them.” — Leonardo da Vinci

Building a deep tech company from scratch for the second time, I am more convinced than ever that the above title is true. There is a general perception that deep tech is expensive and it takes a significant amount of time. While that might be true in some cases, it is not hard wired into the rules of deep tech. This perception originates from the fact that the process of developing, refining, debugging and commercializing deep tech does require a relatively large number of complicated iterations. By iteration, I mean repeated experiments and trials, until one gets to the optimum product. Since deep tech products have a myriad of complex components working together, even a small change can have a cascading effect on some or every component. This adds to the complexity of the iterations. But this is also where things start to get interesting–if these iterations are delayed AND too few lessons are learned from them–the total number of iterations needed to get to the final product increases. This makes a project time and cost intensive.

Let me explain why. Each iteration/trial/experiment has a capital cost to it–money spent on buying raw materials and any new components for the iteration. Let’s take a simple example. Say you want to change the chassis of a car from steel to new aluminum alloy, lighter but strong as steel. It would require purchasing the new alloy, and the components needed to process it into the right shape. Then there are variable costs such as facilities, utilities, salary, etc. needed to install the new components and test it. Variable cost is a function of time that mounts significantly when a trial takes longer. The variable cost in most cases outweighs the capital cost because of the sophistication required in the equipment, facilities, and the need for highly skilled personnel, all of which are expensive. In this scheme, as the iterations increase in length and number, the variable cost increases and thus the total project cost exceeds expectations. I define Execution Efficiency Index (EEI), which is the ratio of capital cost to the total project cost and gives insights into how well a team is executing. My personal experience with EEIs have been between 0.1 to 0.3 but can vary widely depending on the project at hand. EEI can never be 1 and the goal is to pull it as close to 1 as possible.

I have seen two ways of building a deep tech company. The first way is to raise significant capital by breaking the technology development process into sellable milestones for each funding round. Or the second way by being extremely focused on extracting all possible lessons from iterations, while reducing the cost of each iteration. We prefer the second approach. Of course, this is easier said than done. It requires excruciating focus and a team that is obsessed with the problem, meaning they think about the problem even when they are not working. In addition, starting a new company allows for some creative levers to be pulled depending on the nature of the technology. As an example, in my first company we were casting and forging new alloys, modeling industrial furnaces, and scaling material manufacturing for fortune 500 companies. We were undergraduates with no access to any real capital or the infrastructure needed. Our creative solution meant striking a deal with a metal producing company. We provided free services on their technical challenges, and in return we asked for R&D space and access to their industrial equipment. We saved rent, some equipment cost and had access to industrial manpower time, all of which got us started without significant upfront cost.

Cache is founded on the principles of frugality and speed. We optimize for the speed at which we can get to a reliable energy storage system followed by what is the most economical path to get there. The team appreciates that building a breakthrough technology isn’t enough, we need to run the system reliably over decades and earn the trust of our customers. Speed and frugality is not just about spending less, but having abundance to do more and channel resources to critical bottlenecks in the path of technology adoption.