When we look across the world for E2W success stories, nothing comes close to India-based Ather Energy. From starting as a college-incubated startup (IIT Chennai – no small feat!) with a market share of more than 17%, to listing on the Indian bourses with a present market cap of USD 3.2 billion, as we write this, Ather has had the most complete journey for a startup. Today, they make what are probably the best electric commuter scooters in the world. They are the perfect template for any E2W startup to emulate.
It was our delight to sit down with Swapnil Jain, the CTO & Co-founder of Ather Energy, for an InsightEV interview. When we did that, the focus was on taking away learnings from Ather’s journey for a new crop of mobility startups across the world. Swapnil is humble and super-confident, and did not hold back in delivering nuggets of wisdom that are gold for anyone planning to walk the path.
One of the things I’ve noticed about EV startups is that their initial teams are very small. Ather has been through their journey, and I’m sure you started with a handful of engineers working with you, creating the first product. From there to today, there has been this incredible transformation in the team and systems. I would like to understand that part of the journey. What changes from the team perspective, from a small team to a large team, as the company evolves? What is the likable thing for you as an engineer? What is not a likable thing for you as an engineer?
Swapnil: The likable bit is that as the company grows, you have more work to do. Second, you have more specialisation requirements. You realise that while a particular engineer knows something, they may not know other aspects. Like, a design engineer knows about design, but they don’t know enough about manufacturing, or they don’t know enough about supply chain.
When you start adding more people to the team, you start adding specialists. You may have a mechanical engineer, but you also need a materials engineer. So you add, and start realising that over a period of time, there is a lot of specialised vision, which is required, and that’s a part of it. When you have more people who can give you more perspective, it allows you to solve harder problems, because the reason you can’t solve a problem is that you don’t understand it well enough, and you can’t solve it. The specialists that you hire give you the possibility to be able to solve a lot of harder problems as well. It allows people to focus on what they are good at and not waste time on something that they may not be very good at.
But the same thing brings in problems where ownership starts diluting, because now everybody is only owning a piece of the puzzle. The end-to-end ownership becomes sort of questionable. Similarly, people lose context because they are only looking at a small piece of the problem. Everybody has only half the information. Nobody has full information, and sometimes, to make certain decisions, you need full information; you might make a sub-optimal decision if you are only working with half the information. That’s where it starts getting challenging.
People are good at things. Sometimes they are attracted to the company’s culture, but the problem is that the culture also starts diluting as more people come into the picture. Automatically, you start taking less ownership. If I have to work with five people to get something done, I do it as a job. But if I am the only person who’s been made responsible for doing it, I will have the full responsibility to make it work, and I will get it done somehow. I think that’s where the biggest challenge comes in.
So, taking a step back from this, if today there is an electric mobility startup working with a small team, what would be your advice to them? From a team perspective?
From a team perspective, for now, I’ll come to the technology a little later. From the team perspective, I think building at any point in time, you have to keep relooking at your org structure as you evolve, because this problem of zooming in and zooming out is very real. The more people you have, you can add a layer to be able to stitch across those things.
So if I had only one battery engineer, I could give only one engineer. But if I have more, then the work can be more specialised – like someone to do the mechanical, somebody to do the electronics, somebody to do the software, and somewhere up a person to hold the entire piece together because a battery is a battery. It doesn’t care about the various parts and the software inside. It’s a single piece.
So as the org grows, you have to add layers to keep stitching things. It’s very much like building a country. You have to look at product development also in a similar manner.
One of the finer nuances, and one more thing I would advise, is that when you start, you should not have too many specialists because they can’t get anything done by themselves. Generalists are people. who can often be sub-optimal, but will get you started. I think in the early days, you need people who can do that. They may not be the most optimal, but they will get you going. As the org grows, you can add more specialisation. If you add too much specialisation at the start, you’ll not get anything done.
That’s very interesting, because I would have thought the other way, that you need to have a lot of specialists. After all, there are specialised components like batteries and electronics.
Yes, because a battery has set dynamics, and there’s BMS, there’s an algorithm in the BMS, there’s firmware in the BMS, and there is hardware in the BMS, and then there is structural integrity of the battery. Also, the thermal management of the battery. These are all different domains, and people do PhDs in each of these areas. So, if you have a person who is very good at cells but doesn’t understand mechanical engineering, you will not be able to build a battery pack.
But instead, you can have a generalist who may not be an expert in any of these, but is good enough at least on a basic level, understands a cell, understands a battery, and understands at least four out of five things, then that is the right person for you to have. That is the person with whom to build. They may not be the best to begin with, but at least they’ll be able to build a battery, and they will be able to understand which areas they are lacking and then bring in those sorts of specialists.
That’s what has worked for us in the early days when we had a lot of engineers who came from Formula-C. The biggest advantage is that they just don’t understand theory, but have also built things with their own hands. So with the same engineer, I could let them design the part and also trust them to go and get prototypes and build the vehicle and tighten the bolt whenever bad. Otherwise, I would need to have a mechanic who understands some things but doesn’t understand how to get prototyping done. Then others may not be hands-on enough to be able to tighten the board. So these are the things that are really important. You need more generalists at the start and more specialists as you get going.
One of the preconceived theories about electric mobility is that it is easier to do, and that’s why you see a lot of startups. Let’s say, designing and manufacturing a 125 cc engine needs large teams, a lot of metallurgy, and components that you can’t really build in a shed. However, electric mobility is considered easier. But when you actually go from the stage where you are tinkering with things, to get moving, to a stage when there is something that is ready to be dispatched to the customer, people miss the journey between tinkering and dispatching. Can you elaborate on that part? What complexity comes with that?
I think that it’s very easy to tinker, easy because there are fewer mechanical components, easy to put them together, and make them work. But it is equally complex because you can’t see a lot of things. You can’t see the current flowing. You can’t see the messages, CAN messages going between the two devices. You can’t see the voltage buildup, which is happening. And the same thing that made it easier now has made it very difficult because a lot of the things are either non-intuitive or are not visible to you. That’s where things start falling apart for people because a simple root cause can cause a lot of problems. They can’t even see what is failing.
That’s where the difference between a prototype and the product dispatched to the field is, because you have to think about so many things that can go wrong and design for it and test for it. This can be a challenge if you are only used to tinkering and don’t understand the design and manufacturing aspects. That’s where the specialism part starts coming in. If you are a generalist, you can tinker, but you will not be able to think about those corner cases which happen in the corner, not only in engineering, but also in manufacturing.
People often tend not to take into account process variations. I mean, the design engineer will never worry about process variation, but when it comes to production, that is the most important thing you have to worry about. Process variation will make or break your product. You have to think about your assembly line, something that an engineer can assemble. The technician is not going to assemble in the same way.
Similarly, you have to worry about so many corner cases, which will probably happen once in six months, but you have to worry about them. It may happen only for 10% of the fleet, but you will have to worry about them. A lot of these are things that either you have seen in the past, or you will understand them as field data starts coming in.
So having strong first principles is also important. If you know first principles, a lot of these issues you can anticipate in the strategy and take care of them. If you acknowledge more of these things, you will anticipate and take care of them. But for some of these things, you will just wait for the field data to come in and learn about them.
I’ve always wondered, the old world motorcyclists were all about how the machine handles corners, accelerates, and has a top speed. We know that there are things that electric two-wheelers do very well, and then there are limitations. I look at it as one world moving towards another world. There has to be an evolution in the customer, too. But how comfortable are you about what electric can do, what it has been doing for some time, and where you see the limitations or the challenges are for electric machines to match the ICE experience?
So I think if I were to specifically talk about something like a scooter, I think there is very little that today an EV cannot do, which an ICE can do, other than the fact that refuelling is not as instantaneous. I actually feel, again, for a scooter customer, that it’s an advantage to be able to charge at home, rather than having to go to a petrol bunk every few days. But other than that, I think in the scooter world, there is no very strong reason for somebody not to convert to an EV.
It is motorcycles where some of the challenges start coming in because they have a lot of emotions to it. There is an element of machismo. There is an element of sound that you get in the product, and sometimes even long-distance riding capabilities. Then again, within the bikes, there are certain areas like off-roading. I think EVs are best for off-roading. You get that low-end torque. You have a lot more control over your torque. You don’t need gears and all of that, and actually, that serves as an advantage in the off-roading world.
I think even for a city, they’re pretty good EV bikes. To some extent, for some people, they will miss the thump of the ICE. I think that’s not something that I’d try to compete upon. There will be people who would want them, and for them, probably ICE is still okay. But there is a large majority for whom the zippiness of the EV, plus the convenience of the EV, is a strong point.
I actually think about ride and handling, and EVs can match or do better than ICE. So if you’re a true-blue performance lover, then I don’t think you would have a lot of problems with EVs. But if you’re only there for the emotion, which is how it sounds, that is the only area I would not even compete in is what I would say.
Is the weight delta something that you think will go away between ICE and EVs?
That’s where I think the problem of long range comes in. If you want to do a long range, you have to carry big batteries, which makes it heavier. And I think that is harder to do. I mean, again, it depends. If you’re cruising at like 70-80 kph, then I probably give you some amount of advantage for long distances. But if you’re really trying to do like 100-150 of sorts, then it’s very hard for EVs because on ICE, even if you run out of fuel, you can at least gas it up quickly. EVs will struggle to. I mean, you can get it, but is it really worth solving for today? Maybe, maybe not. The way the delta is, and it comes because of the battery, trying to beef it up by adding 10-12 kWh of battery. It makes it impractical for the city, so it may not be a good product, which is what I feel. A better product choice is to build something with a less than 10 kWh pack.
Again, that’s my hunch. I mean, we will still want to experiment with many things, but probably 10kWh max is what you look at, but then definitely there will be a weight delta, which will come in. So I think weight is harder to get rid of. You need the range, you need the battery. You can’t do much more.
I noticed that most Indian manufacturers operated at 48V. I think Ather is also at 48. I know this was a conscious choice at the start of evolution. But now with people moving to higher voltage, do you see that as an attractive option?
For bikes, definitely, because you need more power levels, it will make sense to go to a higher voltage platform. Scooters, we believe, 48V is very good. You only unnecessarily increase the cost of the components by going higher, and other than the 450X or an Apex, I don’t think any other scooter really requires a higher voltage platform.
As Ather evolved and you now have the new Rizta platform. The Rizta now contributes to 70% of sales, while the 450-series is 30%. The Rizta is targeted at the family, and your new EL01 concept is also a family scooter. Well, 450 is an all-out performance scooter, and there is an enthusiastic crowd that looks at the 450 as the benchmark of performance scooters, entertaining performance scooters in India. Do you see Ather moving away from that now that there’s a lot of family-oriented products in the showroom and in the pipeline?
No reason for us to move away from performance. It was always known that it was going to be a smaller part of the portfolio. But then, there is a market for it, and that’s why even with the EL platform, we’re trying to build a platform on which we can build multiple products. serving the different needs of customers. So, yeah, it will be a smaller set of people who will need it, but I don’t think that with electric, because we get so many possibilities in terms of performance, is there a need to go a lot beyond what the 450 does? Maybe not so much.
Do you want to go on a scooter beyond 100 kph? Maybe not.
So maybe not a lot of change in the performance levels. As products, you would want to build more performance products. Not higher-performance products, but probably solving different problems.
You know there’s a certain enthusiastic twitchiness in an electric vehicle. You have high torque, and it takes off faster. As you do more family, and I also call it slightly higher-age focused products, do you see this twitchiness intentionally being toned down to appeal to a wider audience? Not everyone can handle that kind of acceleration, especially a family man.
I mean, that’s what we did with our Rizta itself. We were naive at the start, when we thought that why would people not want performance? You can always control it with the throttle. But we later realised that people just want to be sure that even if I pull the throttle completely, I’m not going to be in a difficult place. That’s what we did with the Rizta. We actually toned down the performance, even though it was not a major cost advantage, but we realised that the customer wants us to tune down the performance. I think that’s a real benefit for the people who want vehicles, which are not always about performance. So I think it’s good to be a bit slower, a bit less acceleration is what you would need. But otherwise, the Rizta is almost ballpark there.
Ather has always done scooters with fixed batteries. The battery is pretty much an integral part of your frame. Do you ever see yourself doing a product with portable batteries in the near future? Is that something that has crossed your mind, something that has been discussed internally? Why is your focus on fixed batteries?
Yeah, we don’t. The operations aspect of handling a swap station is a big pain. It adds a lot of cost, which is not visible on day one, starting from owning the space, owning the extra set of batteries, to transporting the batteries to different locations, plus, let’s say, if you have an operator, and then the charge station costs. There’s a lot of cost that comes with a swappable battery, which people are not able to just give. We have seen a lot of players come in and go.
Unless you’re doing a fleet. If you’re doing a fleet, definitely you should do swapping because your agenda is to sweat the asset a lot, especially if you have your own fleet, and then I would say, you should not even try to think about charging. You should only do swapping. That gives you a very good benefit.
But if you’re not doing that, if you’re a regular customer, the convenience of charging is more. The other option could be okay, I can detach the battery and take it to my house and charge it. Again, I want to get very practical. Nobody wants to carry batteries all the way up to their floor in a lift. So it sounds like it solves the problem, but when you look at it deeper, it’s not very practical to do that.
Hence, we have (fixed batteries). It’s not a philosophical call; it is a very practical call, and it just doesn’t make product sense to build as swappable. But we’ve seen that multiple times. Anybody who came up with a swappable battery just couldn’t justify it over a period of time. Most of them fail, even in setting up the station itself.
The other thing is, if you work with somebody else who’s providing a swap-station, but then you have to work with their design and their own batteries. The most important part of the vehicle is the battery, and if you lose control of that, how can you promise a certain experience to the customer? If you’re keeping the batteries with yourself, there’s no way you can really scale this business up.
One of the experiences that startups go through is that there are a lot of things that are available off the shelf. You start doing electric mobility, and you can buy complete packs along with BMS, motors, and controllers. Actually, you can put things together and assemble them. Looking at your journey, what would be your advice to startups? What should they buy off the shelf, what to develop in-house, and what to outsource?
It’s a lot about what is core to you? So first is what is core to you, and you may have a strategic aspect to it. The second is customer experience. So if something is very core to your customer experience, or to the product itself, and if that expertise is not really available outside, then you should build it yourself. Because what we have seen is that, sometimes, there will be somebody who says, Hey, I have made the VCU, so I will also make BMS. It doesn’t work that way.
Yes, you may have experience in making hardware, but a BMS is a very different ball game, and they’re going to learn at your cost anyway. Plus, once they learn, they’re going to supply it to everybody. So you don’t want to be in that place because you are the guinea pig in that area, and you don’t want to be anybody’s guinea pig. If you learn something, you want to keep it to yourself. So if the skill doesn’t exist outside and it is core to a product experience, then you should do it in-house.
If the skill exists, and it is probably not critical to your product experience, you definitely buy it from outside. There’s no point trying to build your own suspension. The skill of building a suspension is out there. There’s no point trying to build a tire or a brake system. There’s nothing new that you’re doing.
So that’s the third dimension. Is there any novelty in it? If there is no novelty, if everybody is building brakes, you just buy brakes off-the-shelf. So, the only place where I would be really careful about is that while some of these things you might be able to buy independent (and one of the reasons why a lot of four wheeler OEMs failed in some of these transitions is because), one of the disadvantages which comes when you buy things from outside is that they optimise it for themselves and for multiple products. So if I’m a, let’s say, a battery supplier, I’m not going to optimise it only for an Ather. I am going to make a standard so that I can minimise my development costs and sell it to everybody in the industry. Similarly, a motor controller guy is going to do the same thing, which essentially means that when you put these two together, they’re not the most optimal.
They are suboptimal because they are also developed for a large pool of people who have very different requirements. The other one is also developed for a large set of people who have very different requirements. And now you have a suboptimal product because it’s the lowest common denominator, which you have received and are paying for, yet you’re probably using the least of its features. That makes the product very beefed up unnecessarily in terms of the requirements that are there. It doesn’t give you the most optimal experience.
This is also the reason why we decided to do a lot of things in-house. When you realise that when I take something from vendor A and a vendor B, and put them together, I don’t get what I want. I am not getting the thermal performance, which I want. So if I want to build a product, it’s just an amalgamation of multiple specs; it does not give me the experience. The core product experience is getting disturbed if I’m just buying things and putting them together.
So you have to understand what is core to your product: maybe one or two things you can do in-house, and if the skills exist outside, maybe pick up one or two things from outside. But you have to build the core yourself. Maybe over a period of time, it is okay to standardize. Like what happened to the ICE industry over a period of time, everything is sort of standardised. Almost everybody in the industry uses similar kinds of things. Probably you can just work with somebody else. Or if they have developed a better skill, then probably you can work with them. Otherwise, you should not let somebody else learn at your cost. That’s the simplest thing that I would.
In the case of Ather, what do you do in-house, and what do you decide to outsource? I think you do all the hardware in-house?
Everything that is core to EVs and the software, we do in-house: battery, battery management system, motor controller. Only the motor is something that we don’t make in-house. But we are very closely involved in the design of the motor. But other than the motor, full entire battery and the system, the motor controller, algorithms of the motor controller, electrical architecture, the vehicle control unit, the DC-DC converter, yeah, these are the things which we do, even the dashboard, though it’s not core to Ather, but then it becomes core to the software experience because dashboard is HMI for the software experiences.
Then, further down the entire software track from the dashboard to the cloud and the mobile app are things that we do in-house.
When you look at an ICE R&D, there are a lot of mechanical engineers, some electrical, very few electronics, and very few software engineers. While in EVs, the complete system goes through a structural change, and that has been a challenge to the incumbents. Like, how do you do software when you are not a software team? It’s a culture change, and I see that as an advantage for startups. Now, if you were to look at your team, how big is, on a percentage basis, the software team versus the hardware team?
So, software is of different types. If I look at all sorts of software together, it will probably be 40% of the entire engineering. If I look at only the firmware part of it, that would be a very small portion, probably about 10%. But there is firmware, then there is HMI, and there’s the mobile application, the cloud, and then there are also some of the enterprise tools, which we also make in-house. So all of this put together is about 40%. So, 35-40% of our engineering org is software, and that gives us a lot of advantage, because controlling every piece of software is important.
If you outsource a lot of software, then it becomes very hard for you to get ahead. Today, the biggest advantage for us is that tomorrow, if we decide to change our electrical architecture, because the software team is with us, in parallel, we can change the software architecture and get everything together. Otherwise, changing an electrical architecture could become a very big activity. It is difficult to work with multiple players because no single player in the automotive industry has all the skills. The same person will not understand a BMS and a motor controller, a charger, and a DC-DC converter. You rarely find people who can do all of them really well.
In the past, we have done re-architecture of the product at least three times, and one of the biggest reasons we could do it is that the entire team is in-house.
Now that the place that you have reached, you have like 100s of people working in your team on both the software and hardware side. Going back, if you would like to advise a startup on what part of software to do in-house and what part to outsource, because the start-ups have 10-15-person teams, and it is very difficult for them to do everything. What would be your advice to them?
If you can find somebody else who’s doing this better than you and it’s not a part of your core product experience, you should definitely outsource. For example, in our case, even today, we are not working on the OS Platform (Android). We are not trying to do anything on the cloud infrastructure. That we buy. So there are these aspects that, irrespective of whether it’s a small team or a large team, you should not do.
But at least in the EV space, it’s very difficult because hardware and software come together. So you can’t say, okay, I will do the hardware myself, but do the software outside. It doesn’t work really well. Even within the organisation, it’s hard to get these two teams to work together to deliver it. So if you’re two different organisations, it is definitely not going to work.
So I think, whatever is core, you have to do it. You know, it is difficult, but it is very critical. Today, you are not in a position where you can easily get any of these software stacks from outside. Hardware may be easier, but the software stack is difficult to get.
Cloud, again, infrastructure level exists, but applications you have to write yourself. You should not try to do anything on the infrastructure side, definitely. Thankfully, some of the IoT stacks have become a bit more easily available, but still, you have to do a lot of modifications for it to make it specific to your use case.
