TRIANGLE
BLOG

If you haven’t read my previous blog post ("Saving the European Automotive Industry"), I’d recommend reading that first, as it sets out the “why” for building a car. It’s important to outline that as there have been many automotive startups that have tried to build a car and failed. In fact, if you look at the statistics, creating an automotive startup might be one of the dumbest things to do.

 

So, it begs the question as to why these companies fail and what we are doing differently:

​

1. Most startup automotive companies try to compete with legacy manufacturers by trying to be more luxurious, higher performance, better this, better that. However, who in their right mind would drop a very significant proportion of their income on something unproven, from a company that most people have never heard of, for a slight increase in performance or any one of the other vain metrics? The answer is very, very few people. What this means is you have a hard time gaining adoption to cross over from the brave early adopters to mainstream consumers.

​​

2. All of the above assumes you actually get to a product in the first place. Another high percentage of companies die in the process of even getting to a prototype, never mind an actual production version. The critical problem here is that a car is an insanely complex system, with multiple elements that have to be intertwined, and oftentimes the people starting these startups don’t understand the vehicle as a whole system and have surface knowledge on a lot of the critical elements. I’d like to point out a stark observation: if we look at the successful automotive startups within the past decade, take Rimac, Rivian, and of course Tesla—they all have a founder who was deeply invested in every aspect of the vehicle and understood it intuitively. So much so that the founders of Rimac and Rivian built their first vehicles either single-handedly or with very small teams. Therefore, the founder must be someone who can almost single-handedly design and put the initial vehicle together. Without this level of engineering detail and understanding, it is very difficult to align the company towards a product that can even remotely compete with competitors.

​

3. Assuming you do get to a compelling prototype vehicle, the next challenge is mass production.

​

 

“Manufacturing is insanely difficult. It's under appreciated in its difficulty.”

Elon Musk

​

 

This requires two things: firstly, the design needs to have been thought out in a way that is compatible with manufacturing at scale. This includes a huge number of variables from the material selection, the manufacturing process, and then finally, ease of assembly. Oftentimes this means you pretty much have to redesign the vehicle almost entirely again, while simultaneously designing the actual manufacturing line as well. Some companies decide to outsource this and subcontract out the parts, while aiming to do the final vehicle assembly in-house. However, without the order quantity and purchasing power of a large OEM, this results in extremely inflated vehicle costs. From talking with engineers at Arrival, this was in fact their failure mode.

​

Companies also fail due to the extreme amount of capital required to set up a vehicle production line. This is mainly because smaller companies try to use the same manufacturing techniques as an OEM that needs to produce thousands of cars a month. This means you have to invest in all the tooling, robotic manufacture, and scaled production methodologies, while not having OEM levels of orders. I’m still unsure as to why this happens, but it may be that scale-ups employ engineers from large OEMs, and therefore they have only imagined scaled vehicle production in a very specific way.

​

So, what are we doing differently?

​

1. Firstly, we are not competing in the same market as our competitors with better performance or luxury (although I’m sure most people will be surprised to see what we come out with). But by solely focusing on producing what our customers want at radically reduced cost, by a minimum factor of two. Only then do we believe we can drive significant market share and also compete with Chinese EVs. Does this mean terrible vehicle quality and cheap materials? Absolutely not. In fact, the raw material cost within a vehicle is actually a relatively small percentage (around 10%), and by using high-quality raw materials, it does not significantly increase the vehicle cost. The trick is in how you convert the raw materials into the shapes you want. (More on that shortly).

​

2. Now, as an early-stage company, we are not going to have the design resources to produce a huge number of prototypes in order to get to a compelling design. Nor do we have thousands of engineers to tackle every aspect of the vehicle. Here I have a slightly controversial view: I actually believe that a team of no more than 10 people is required to design a compelling vehicle. This may sound crazy as most OEMs have hundreds of people working on a single design. But this is a circular problem: after a certain point, the more people you add, the efficiency of the entire team drops faster than the gain in engineering resources.

 

There is a concept in software engineering of the 10x engineer, and this is also true in hardware. In fact, in reality, it may be closer to 15-20x in the world of atoms. The reason for this is that the more of the problem a single engineer has in their head, the easier it is for them to make design decisions holistically. And in the hardware world, there are so many microscopic decisions that need to be made day in, day out, that a small focused team of 10 20x engineers will easily outperform a team of 100+ engineers.

 

In fact, you can see this for yourself, as there have been numerous examples where people in their garages have built entire cars from scratch themselves, just as a weekend hobby. My favourite at the moment is Josh from Motorsport Engineering, who has pretty much designed and built an OEM-quality race car himself.

​

However, at Triangle, I believe we need to go even further. Is there a way to make a 20x engineer a 100x one? I won’t go more into that in this post, but I’ll be making a post on this shortly. Essentially, what that means is, from a design perspective, I believe we can get to a very compelling vehicle prototype for relatively little capital expenditure. In fact, I think we can get to a prototype in under $2 million, roughly 10-20x less than a typical car company, using a very focused team of the world’s best engineers coupled with an AI system.

​

3. Finally, mass manufacture—easily the hardest part of this endeavour. What we need to do is think completely in the opposite direction of what a typical OEM would do. The capital required to pull that off is not practical for a startup to deploy and would be an incredibly risky strategy if you had to raise hundreds of millions of dollars in order to start scaling.

 

Now, I did say in point 1 that we were solely focusing on how to reduce vehicle cost, but that is not quite the whole picture. It is how to do that while also minimising capital expenditure to scale production. What that means is that we actually end up with a slightly more expensive vehicle to produce per unit, but the capital expenditure to get there is an order of magnitude less. Thereby, drastically reducing the risk of not being able to raise the required capital for scaled manufacture.

 

How we plan to do this will define whether Triangle is a success or not. What I can say for now is that this production line will be something special if we can pull it off, and it will really unlock low-capital scaled manufacture of not just vehicles but of anything you can think of later down the line. (It’s not 3D printing, by the way—I’m against that for a number of reasons!)

​

Overall If I am totally honest, the chance of success for Triangle is around 20%. However, this company has to exist. Without it the European automotive industry will die. Unfortunately I have seen close to 0 action by any of the big players to get their costs down, which is a necessity in order to compete with the Chinese. And so either we can wait for their demise or do something about it. Hopefully if we show what is possible with Triangle, it will catalyse the industry to follow suit.

​

Lets build!