Have you seen the TV show? Aspiring entrepreneurs pitch their ideas to “shark” venture capitalists. Well, here’s how young Tesla Motors conceived, polished and pitched its big idea, as recounted by co-founder Marc Tarpenning in this under-exposed interview. TeslaMondo has mildly refined it for logical and temporal flow, and to delete the words “sort of,” which Tarpenning says almost as often as JB Straubel says “you know.”
This is a long post for a site that favors brevity, but this is the stuff of a good book. Maybe even a good movie. Certainly a good episode of Shark Tank, if TV allowed for a longer format. This ain’t no five-minute sales pitch.
Context: Tesla’s co-founders, Tarpenning and Martin Eberhard, had just built and sold an ebook specialty company called NuvoMedia, and now had cars on the brain. They had zero automotive experience, but that would prove beneficial.
So, Mr Tarpenning, the microphone is yours:
Martin had this idea of — he’d wanted to buy an electric car, and it was right about the time the zero emissions mandate got re-written in Sacramento, so all the electric cars had ceased to exist. And they weren’t very good anyway. The EV1 was the only one that was kind of viable. And it was all-lease. And of course, GM took them back and crushed them.
So he was pissed off about that, and he said, ‘Why don’t we do this electric car thing? It can’t be that hard.’ And I thought: How are we ever going to make a car? Our experience with batteries in the consumer electronics world, with the ebooks, we knew batteries were getting better and better and better. They had gone really from nicad to nickel metal hydride, and then lithium-ion . . . so we were sure with some spreadsheet calculations that, in fact, you could put enough energy on a car to make it really compelling, I mean way better than any of these EV1s or anything else they’d ever imagined, and the rest of it is a computer science problem — because it’s really just computers synthesizing waveforms to make motors go — and a network problem because there’s lots of little computers required to do all of that, but, you know, that’s our world. Little computers that control stuff? Silicon Valley is really good at that.
Fuel cells needed debunkery right away.
If your goal is to reduce energy consumption — specifically oil consumption, but in any given resource — you want to use it as efficiently as possible, right? So you don’t want to pick something that consumes lots of it for no apparent reason, and hydrogen is uniquely bad . . . It’s a scam as far as I can tell, because the energy equation is terrible. It is just terrible.
The proponents always say, ‘Oh, hydrogen is the most abundant element in the universe.’ But it’s abundant out there in the universe, not here. We live on a planet where hydrogen is super reactive. It’s bound up into everything. It’s bound up into water, wood and everything else. They only way that you get hydrogen is that you have to pour energy in to make it. You have to break it free from the chemical bonds it’s bound up into.
Electrolysis is the most common thing. You put electricity in water and it separates it, but you are pouring energy in order to make hydrogen, and then you have to compress it and that takes energy, and then you have to transport it to wherever you actually need it, which is very difficult because hydrogen is actually much harder to work with than gasoline or even natural gas — and natural gas is not that easy.
And then you ultimately have to place it into a car where you’ll have a very high-pressure vessel which offers its own safety issues – and then, on top of that, you have to convert it back to electricity to make the car go because hydrogen-powered cars are really electric cars. They just have an extraordinarily bad battery.
Hydrogen is an energy carrier and not a primary fuel source on this planet. Maybe out somewhere in the universe, but not on a terrestrial planet. When you add that all up, it turns out that the amount of energy in per kilometer driven is just terrible. It’s way worse than almost anything else you can come up with. Which I always suspected is one of the reasons why the energy companies have been big proponents of it.
When we were raising money the first time, we had very carefully gone through the math to understand fuel cells because there was a bunch of money going into fuel cells at the time and we also looked at biofuels and ethanols. We sort of went down the whole list to figure out what the most energy efficient system was, which turned out to be battery electric cars.
When we would go to these venture capitalists, we had a whole slide deck on why fuel cells were a bad idea, and about half the VCs would get to those slides and they would say, ‘Oh, skip this. We know it.’ And actually, one of them used the word ‘scam.’ But the other half of the VCs were quiet. Then they asked lots of questions about it, and they’d have us go back over the slides again. Because they had, in fact, invested in fuel cells. All of which went out of business.
Green customers have the green.
GM is not the most well-loved brand among the money people. If you’re successful, you tend to buy a Mercedes or BMW. You’re not going to buy the latest GM vehicle . . . yet the people who were leasing the EV1 in California were in the top one percent. They were clearly not trying to save money. Gas was at about a buck fifty a gallon at the time. They literally were spending more at Starbucks than they were at the gas station, and yet they were leasing electric cars, so they were doing it for some other reason.
Literally every driveway had a Porsche and a Prius. They had traded in a BMW or a Lexus and replaced it with a Prius. They were making a statement. They wanted to be less a part of the oil economy.
We looked at that and said shoot, any new technology is going to be expensive, we’re going to come in at a high point. What extra value can we deliver? And it turns out electric cars are really, really quick off the dime.
The mathematical case.
It’s all high school math and physics, really. You’re trying to figure out how much energy it takes to accelerate something. How much friction you have on the wheels. An estimate of the aero drag based on the typical sports car. You know, we didn’t have the actual design. You could make a pretty good guess at how heavy the vehicle is going to be. The spreadsheet predicted what was possible through the math. That’s just amazing to me.
Rampant automotive outsourcing proved perfect for Tesla.
The more we researched the industry, we discovered that over the previous 30 years, the car industry had gone from being vertically-integrated — the most famous being Ford, which literally would take in iron ore and produce Model Ts at the other end — to the current model, where the car companies simply do final assembly. They make their own engines, but everything else is done by other companies . . . And even some of the final assembly was completely outsourced as well. At that time, all the BMW X3s were being made by a company called Magna Steyr out of Austria. BMW had very little to do with it. Magna Steyr built all the cars. The Saabs were being done by Magna Steyr. A couple of the other American car companies that had European versions, Magna was building. So we knew it was possible to go to a company like Magna and say, ‘Ok, here’s the drivetrain, and the suppliers are going to be shipping you parts. Can you screw the thing together?’
Magna was too “magna” to bother with little Tesla. Enter Lotus.
Lotus had experience making low-volume sports cars, and had experience doing outsource manufacturing. They had done this for GM. They had made the Opel Speedster. And a Vauxhall something-or-other. So they had a legal structure that could accommodate us. And they had this idea on the production line, that some of the cars they were making were Lotus cars, and then some for other companies.
But little Tesla had a large-scale appetite for batteries.
At the time, every laptop had these 18650s, and every camcorder. We were familiar with them from the ebook days. We thought, ‘Let’s use these.’ I thought that was insane. There were going to be thousands of them. But the more we thought about them, this was a great thing because they were commodities, made my dozens of different companies around the world — not all of them the same quality, but the same form factor — and that’s what you want. If you’re a buyer, you want to have lots of choices. You don’t want to be locked into one particular weird format that only one company makes. You want to have lots of competition all beavering away to make those 18650s cheaper and better for you. So the whole chicken-and-egg problem around specialized automotive batteries goes away, because we’re just buying these laptop batteries.
Now, it turns out that laptop batteries have to be handled very carefully when you combine a whole bunch of them together. And the battery companies really, really didn’t want us to do that. They did not want us to put more than seven of them together, because there is a chance of a fire. If you mistreat then, you can get them to catch fire. They don’t explode, but they kind of catch fire . . . Sony quite famously had a whole set of laptops that would spontaneously burst into flames.
The Roadster would require about seven thousand per car.
So battery companies really didn’t want us to do that. But eventually we convinced them that we know more about how to keep batteries safe than they did, and so we got a supply agreement worked out. It turns out of you go to sales guy and ask, ‘What is a fantasy customer for your batteries?’ The answer is about seven batteries per user.
And then say, ‘We think your market is a thousand times larger.’ If you can convince even a low-level salesman that their addressable market is really a thousand times larger, you will get a call from the CEO of that company, no matter how big the company is, asking, ‘What’s this thing about a thousand times larger market?’ So we eventually convinced several battery companies to work with us. It helps that we had by far the best safety data. Because nobody else had really tried it.
Ready to pitch the Shark Tank? No. First, some dress rehearsals.
We used some VCs who were friendly to us, but knew they wouldn’t invest in us. For example, their fund only invested in internet routing, or biotech. So we know we were completely outside of their investment space. But we called them up and said, ‘Hey, we have this crazy idea, and we want to pitch you guys. This isn’t a pitch that you’re going to have to actually respond to, but we’d like the feedback.’ It turns out that if every day, you hear nothing but pitches about internet routing, if someone says they’re got this crazy idea for high-performance electric sports cars, all the partners will show up if you provide lunch.
So they gave great feedback, asked quite insightful questions. We thought we had all our ducks in a row and all of our answers figured out, but they would ask these questions we didn’t have answers to. So we would go back and re-think it and re-tool. And that’s how we honed our pitch before we actually used our silver bullets with the VCs that we thought might fund us, because you only get one chance, really.
Smaller funds first.
Unanimity in a big firm is so hard to achieve. Smaller funds? A bit easier. We figured we could probably get a couple people to say yes. We lined up money from SDL Ventures and Compass Technology Partners because they were relatively small funds with only one or two deciding partners. And of course, various friends and family.
We pitched Elon as a super-angel. We went down to SpaceX and did the pitch. SpaceX was just a little startup, years from its first launch. We pitched on a Thursday or something. Then I was in Washington DC all weekend. I was peppered with questions by Elon over the weekend and then into the beginning of the next week. After a few more days: He said, ‘I like this. I’m really concerned about oil consumption as well. I get the vision, and where you’re coming into the market. I’m in.’ It was exactly up his alley.
Raising money in little chunks is the way to go.
Let’s say you need 60 million for some project. It’s not even really desirable to raise that money all at once, because if you’re just a couple of guys with a PowerPoint, your company isn’t going to be worth very much, so to take 60 or 70 million all at once, even if you could get somebody to agree, they’re going to own the whole company . . .
So what you do is, you raise money in tranches. The first in our case revolved around the battery. Because that was really the thing everyone always asked about. ‘Can you really get this kind of energy density in a battery like that, and can you really make it safe?’ So that was the goal of the first tranche of money. The deliverable was a mule, a vehicle that doesn’t reproduce — a rolling test platform.
We weren’t worried about the motor or inverter. Nikola Tesla invented AC induction motors 100 years ago, so we know that was going to be possible. The thing that was a little bit unclear was whether we could really retire the risk around the battery pack.
The best fundraiser? The car itself.
By and large, the VC community is not hugely imaginative. So when you come with a demo that really shows off what you’re doing, it helps investors get comfortable with the idea. So, in this case, we had this little yellow Lotus, and you got into it and it took off incredibly quickly, and it was silent and beautiful. Literally we got working at midnight, and the next morning we had a board meeting. The board is there. Elon is there. We show it off. It was such a rush to ride in that thing. It was such a visceral experience. So then it was easier to raise money.
We had some meetings with other VCs at that time. They came in from Chicago. They saw the place. The heard the song and dance. They all rode in the car. There were only three or four partners. And they said, ‘Hey, can we borrow your conference room for a few minutes? We have a meeting we have to do.’ They had to be on a plane back to Chicago. They come out of the meeting about 20 minutes later and they said, ‘You know, we’re in. That was our partnership meeting.’ And it was because they had ridden in it. It proved to them that this was possible, and it was unlike anything they’d ever been in. This wasn’t some golf cart.
Crash testing, durability testing.
This part of the interview drags on a bit, so TeslaMondo will skip ahead. Basically, Tesla took advantage of various third-party crash test modelers and durability testers. These companies did not exist in prior years, when auto companies did everything themselves. Once again, the outsourcing of the auto industry proved vital for little Tesla.
The test cars had a two-speed manual transmission, which we didn’t think would be hard to source. The car industry was making manual transmissions for a hundred years. So we contracted with a small transmission house. Now, we had a lot more torque than they were used to, but we figured, ‘How hard is that?’ It was not something we really focused on because we had a supplier that we thought could deliver. But their transmissions just didn’t work. They couldn’t shift, which was unfortunate for us.
We then contracted with another, a much bigger auto supplier to build transmissions, and we spent lots and lots of money with them. We looked pretty good initially, but we put them on the durability track, we put them on our dynos, our test fixtures in the lab — and they all begin to break. And they all break in different ways. That’s not good. If they all break in one way, you can deal with it. But they’re really not breaking well.
By this point, everything else is done. We’ve gotten through the federal motor vehicle safety standards. We are essentially ready to ship, except we don’t have transmissions that work. This nearly killed the company. It was also in 2007, financing was starting to get a little skittish, and here was this huge — it was the first milestone that we missed in a big way. Everything might have been a little bit more expensive, or a little bit late, but there were conscious decisions along the way that did that. We had decided to make the car better, or change the specs. Everything along the way was conscious. This one was just a screw-up. And it just wasn’t in our core competency. We were dealing with one of the largest transmission makers in the world. They should be able to make a transmission that works, right? We didn’t understand how dangerous that was.
The savior? Moore’s Law!
As it turns out, of course Moore’s Law is thankfully ticking along. The state of electronics had improved. We could increase the power we could deliver to the motor . . . and get to a one-speed transmission. Which is just a reduction.
Our new supplier was awesome. In record speed, produced a beautiful transmission, and that’s what we ended up shipping. We shipped late. Our customers had been waiting for years.
TeslaMondo note: The first Roadster customers took delivery of cars with the inferior tranny, on agreement that Tesla would upgrade when the new one-speeds became available. That worked out perfectly for all.
A mouse among the mastodons.
We figured that if we were successful, we would change people’s perception of electric cars, that they weren’t these lame golf carts. That people would say, ‘Oh, if I ever get rich, I want one of those electric cars.’
But in terms of the other car companies, we thought they would wake up and say, ‘Wow, this is really, this can be done, these can be really compelling vehicles,’ and there would be some amount of space race, basically, to get into the electric car business, and we would have to deal with that as it went.
And we could deal with that in a couple of ways. We might be suppliers to some of those car companies. We figured we would have more experience, and more electrically-driven miles than anybody else in history, by the time this became a hot thing, so we would be the world’s experts in that, and this would be a good place to be if all the car companies were trying to be in that space, so there would be lots of potential deals, and, who knows, an acquisition? We’d have lots of options. And if the car industry really decided to be competitive with us, we also felt that Silicon Valley could out-compete them, because they just move at a much slower pace.
Big Auto still struggles with car-placency.
I don’t think any of us really imagined that they would move as glacially slow as they really are. I mean, they’re just now, finally beginning to get with the program. And still their technology really just isn’t as good. And, you know, they have huge resources and tremendous experience certainly in making cars. I certainly didn’t expect them to take this long to wake up and begin to take electric cars seriously and begin to produce competitive products — which I think plays very well for Tesla.
If you’ve been around for a long time and you’ve done things the same way, and you’re making money — it’s not like you’re failing — there’s a huge resistance to taking risks. Because that could impact your division’s numbers or your career. It just makes it very hard for innovative ideas to infiltrate these big companies. In Silicon Valley, we don’t have any hundred-year-old companies.
UPDATE: Musk just spent a VERY long time at the shareholder meeting tonight (5/31) waxing nostalgic and telling gallows-humor anecdotes about Tesla’s garage days. It’s a reminder that there’s a lot more to the story than just Tarpenning’s version.