Shahin Farshchi (VC) explains why hardware and robotics startups are riskier than software for venture investors: they have longer development cycles, higher capital requirements, harder/expensive iteration, and require diverse skills and partnerships. He cites semiconductors needing "tens of millions" to reach first product (often sold at a loss initially) and gives Plethora as an example of a capital-efficient approach that uses off-the-shelf robots/CNCs. His practical recommendation: minimize risk by leveraging existing technology and focusing on integration rather than reinventing core hardware.
Chapters
So VCs, as a rule, VCs generally hate multiple dimensions of risk. They like to take risk, risk on a technology, for example, with the expectation that the team can execute on the business and the market will accept this product. They like to take a risk, for example, on a team, but with a proven technology, with a, for example, a huge market. Although the team risk is one that they typically would not like to take. Normally it's a product risk or a business model risk, but not a combination of the two. And typically VCs, especially our firm, really bets on the team. We want the team risk to be minimal, which is when I say team risk, I mean our faith in the founders to be able to execute on this plan and to be able to recruit amazing talent that's smarter than them to help them execute on this plan. So now the question about hardware. What makes hardware difficult? Hardware is difficult for many reasons. First reason is that the time constants or the cycles are long and the amount of money that you need to be able to get to a, whatever you want to call it, the minimum viable product or the first prototype is long. I mean, you can have four Stanford undergrads bang out an app in a week, for example, and they basically use lamp, the LAMP stack, and just get something out there that works pretty well for a minimal amount of money. Good luck doing that. You know, within any aspect of hardware, even if you're soldering stuff together, it'll take a lot of time and it'll take, you know, months perhaps and, you know, many, many thousands of dollars to get you to that first kind of proof of concept and that kind of, and that this challenge basically continues to scale throughout the lifestyle, the life cycle of the company. So if the first prototype took, you know, five times longer and took five times more money, getting it to production will take many times longer and much more money than it would an equivalent, for example, software product. So that's one major challenge. The second major challenge with hardware is that it's very difficult to iterate. If you get your, let's say, if you're making a gripper robot, a gripper device, if you get this gripper qualified, and if you get into production, then if you want to change some aspect of it, then it's going to take time and money to basically be able to make that change because there's a huge fixed cost associated with getting that product to where it is. In software, obviously making changes isn't as challenging. And then in the case of hardware, you need a very kind of diverse skill set and set of, I would say, relationships in order to succeed. Which probably isn't as much the case with the more traditional kind of software companies. The types of people that you have to recruit and hire, the types of people that you have to work with, the people that you have to partner with. It's a very different kind of community. And again, the expense and the time cycles are longer, how much longer and on the average. So, you know, I can give you an example. So in the case of semiconductors, so, you know, I did a couple of, well, we have a few semiconductor companies and it took on the order of tens of millions of dollars for them to get to their first product on the market. And they were not even selling those products profitably. They had, they're selling both. Both the ones that I was very close to were selling at a loss initially until they could scale up and get to the point where they can start selling with margin, let alone be able to sell with healthy margin. So again, the economics are somewhat against you. This doesn't really apply to software companies, which is why you see the vast majority of VCs up and down sand hill putting a lot more emphasis on software. The point was to be mindful, to be mindful of the cost, mindful of the time, mindful of the time it takes to get to a product, time that it takes for that product to get and adoption. And you know, how that would. And how when you take at least venture money, the clock starts and you, you know, you, you and your VCs making money depends on how quickly you can kind of achieve your goals. It's not a matter of if, it's a matter of when. So you know, particular angles. Again to the earlier questions that were asked. You know, if you can make conventional robots, I can give you an example. So I invested in a company called Plethora, right? These guys are based in San Francisco. And what they're doing is making mid scale and small scale kind of manufacturing as economical as large scale manufacturing. So if you want to have a part, you know, at a certain, for example, let's say a buck each, you don't have to wait for three months and deal with a contract manufacturer, whatnot, you can have it done immediately at their facility in San Francisco. And they're using entirely kind of off the shelf robots and off the shelf CNC machines to do the task. So to answer your question, if you can leverage as much as you can, leverage existing technology as much as you can and focus on integration to make it more useful, I think that would be a smart path for a venture backable startup. In my opinion, it.
