by Ashlee Vance
The original plan for producing a prototype vehicle sounded simple. Tesla would take the AC Propulsion tzero powertrain and fit it into the Lotus Elise body. The company had acquired a schematic for an electric motor design and figured it could buy a transmission from a company in the United States or Europe and outsource any other parts from Asia. Tesla’s engineers mostly needed to focus on developing the battery pack systems, wiring the car, and cutting and welding metal as needed to bring everything together. Engineers love to muck around with hardware, and the Tesla team thought of the Roadster as something akin to a car conversion project that could be done with two or three mechanical engineers, and a few assembly people.
The main team of prototype builders consisted of Straubel, Berdichevsky, and David Lyons, a very clever mechanical engineer and employee No. 12. Lyons had about a decade of experience working for Silicon Valley companies and had met Straubel a few years before when the two men struck up a conversation at a 7-Eleven about an electric bike Straubel was riding. Lyons had helped Straubel pay bills by hiring him as a consultant for a company building a device to measure people’s core body temperature. Straubel thought he could return the favor by bringing Lyons on early to such an exciting project. Tesla would benefit in a big way as well. As Berdichevsky put it, “Dave Lyons knew how to get shit done.”
The engineers bought a blue lift for the car and set it up inside the building. They also purchased some machine tools, hand tools, and floodlights to work at night and started to turn the facility into a hotbed of R&D activity. Electrical engineers studied the Lotus’s base-level software to figure out how it tied together the pedals, mechanical apparatus, and the dashboard gauges. The really advanced work took place with the battery pack design. No one had ever tried to combine hundreds of lithium ion batteries in parallel, so Tesla ended up at the cutting edge of the technology.
The engineers started trying to understand how heat would dissipate and current flow would behave across seventy batteries by supergluing them together into groups called bricks. Then ten bricks would be placed together, and the engineers would test various types of air and liquid cooling mechanisms. When the Tesla team had developed a workable battery pack, they stretched the yellow Lotus Elise chassis five inches and lowered the pack with a crane into the back of the car, where its engine would normally be. These efforts began in earnest on October 18, 2004, and, rather remarkably, four months later, on January 27, 2005, an entirely new kind of car had been built by eighteen people. It could even be driven around. Tesla had a board meeting that day, and Musk zipped about in the car. He came away happy enough to keep investing. Musk put in $9 million more as Tesla raised a $13 million funding round. The company now planned to deliver the Roadster to consumers in early 2006.
Once they’d finished building a second car a few months later, the engineers at Tesla decided they needed to face up to a massive potential flaw in their electric vehicle. On July 4, 2005, they were at Eberhard’s house in Woodside celebrating Independence Day and figured it was as good a moment as any to see what happened when the Roadster’s batteries caught on fire. Someone taped twenty of the batteries together, put a heating strip wire into the bundle, and set it off. “It went up like a cluster of bottle rockets,” Lyons said. Instead of twenty batteries, the Roadster would have close to 7,000, and the thought of what an explosion at that scale would be like horrified the engineers. One of the perks of an electric car was meant to be that it moved people away from a flammable liquid like gasoline and the endless explosions that take place in an engine. Rich people were unlikely to pay a high price for something even more dangerous, and the early nightmare scenario for the employees at Tesla was that a rich, famous person would get caught in a fire caused by the car. “It was one of those ‘oh shit’ moments,” Lyons said. “That is when we really sobered up.”
Tesla formed a six-person task force to deal with the battery issue. They were pulled off all other work and given money to begin running experiments. The first explosions started taking place at the Tesla headquarters, where the engineers filmed them in slow motion. Once saner minds prevailed, Tesla moved its explosion research to a blast area behind an electrical substation maintained by the fire department. Blast by blast, the engineers learned a great deal about the inner workings of the batteries. They developed methods for arranging them in ways that would prevent fires spreading from one battery to the next and other techniques for stopping explosions altogether. Thousands of batteries exploded along the way, and the effort was worth it. It was still early days, for sure, but Tesla was on the verge of inventing battery technology that would set it apart from rivals for years to come and would become one of the company’s great advantages.
The early success at building two prototype cars, coupled with Tesla’s engineering breakthroughs around the batteries and other technological pieces, boosted the company’s confidence. It was time to put Tesla’s stamp on the vehicle. “The original plan had been to do the bare minimum we could get away with as far as making the car stylistically different from a Lotus but electric,” said Tarpenning. “Along the way, Elon and the rest of the board said, ‘You only get to do this once. It has to delight the customer, and the Lotus just isn’t good enough to do that.’”
The Elise’s chassis, or base frame, worked fine for Tesla’s engineering purposes. But the body of the car had serious issues in both form and function. The door on the Elise was all of a foot tall, and you were meant to either jump into the car or fall into it, depending on your flexibility and/or dignity. The body also needed to be longer to accommodate Tesla’s battery pack and a trunk. And Tesla preferred to make the Roadster out of carbon fiber instead of fiberglass. On these design points, Musk had a lot of opinion and influence. He wanted a car that Justine could feel comfortable getting into and that had some measure of practicality. Musk made these opinions clear when he visited Tesla for board meetings and design reviews.
Tesla hired a handful of designers to mock up new looks for the Roadster. After settling on a favorite, the company paid to build a quarter-scale model of the vehicle in January 2005 and then a full-scale model in April. This process provided the Tesla executives with yet another revelation of everything that went into making a car. “They wrap this shiny Mylar material around the model and vacuum it, so that you can really see the contours and shine and shadows,” Tarpenning said. The silver model was then turned into a digital rendering that the engineers could manipulate on their computers. A British company took the digital file and used it to create a plastic version of the car called an “aero buck” for aerodynamics testing. “They put it on a boat and shipped it to us, and then we took it to Burning Man,” Tarpenning said, referring to the annual drug-infused art festival held in the Nevada desert.
About a year later, after many tweaks and much work, Tesla had a pencils-down moment. It was May 2006, and the company had grown to a hundred employees. This team built a black version of the Roadster known as EP1, or engineering prototype one. “It was saying, ‘We now think we know what we will build,’” Tarpenning said. “You can feel it. It’s a real car, and it’s very exciting.” The arrival of the EP1 provided a great excuse to show existing investors what their money had bought and to ask for more funds from a wider audience. The venture capitalists were impressed enough to overlook the fact that engineers sometimes had to manually fan the car to cool it down in between test drives and were now starting to grasp Tesla’s long-term potential. Musk once again put money into Tesla—$12 million—and a handful of other investors, including the venture capital firm Draper Fisher Jurvetson, VantagePoint Capital Partners, J.P. Morgan, Compass Technology Partners, Nick Pritzker, Larry Page, and Sergey Brin, joined the $40 million round.*
In July 2006, Tesla decided to tell the world what it had been up to. The company’s engineers had built a red prototype—EP2—to complement the black one, and they both went on display at an event in Santa Clara. The press flocked to the announcement and were quite take
n with what they saw. The Roadsters were gorgeous, two-seater convertibles that could go from zero to 60 in about four seconds. “Until today,” Musk said at the event, “all electric cars have sucked.”6
Celebrities like then-governor Arnold Schwarzenegger and former Disney CEO Michael Eisner showed up at the event, and many of them took test rides in the Roadsters. The vehicles were so fragile that only Straubel and a couple of other trusted hands knew how to run them, and they were swapped out every five minutes to avoid overheating. Tesla revealed that each car would cost about $90,000 and had a range of 250 miles per charge. Thirty people, the company said, had committed to buying a Roadster, including the Google cofounders Brin and Page and a handful of other technology billionaires. Musk promised that a cheaper car—a four-seat, four-door model under $50,000, would arrive in about three years.
Around the time of this event, Tesla made its debut in the New York Times via a mini-profile on the company. Eberhard vowed—optimistically—to begin shipments of the Roadster in the middle of 2007, instead of early 2006 as once planned, and laid out Tesla’s strategy of starting with a high-priced, low-volume product and moving down to more affordable products over time, as underlying technology and manufacturing capabilities advanced. Musk and Eberhard were big believers in this strategy, having seen it play out with a number of electronic devices. “Cellphones, refrigerators, color TV’s, they didn’t start off by making a low-end product for masses,” Eberhard told the paper.7 “They were relatively expensive, for people who could afford it.” While the story was a coup for Tesla, Musk didn’t appreciate being left out of the article entirely. “We tried to emphasize him, and told the reporter about him over and over again, but they weren’t interested in the board of the company,” Tarpenning said. “Elon was furious. He was livid.”
You could understand why Musk might want some of the shine of Tesla to rub off on him. The car had turned into a cause célèbre of the automotive world. Electric vehicles tended to invoke religious overreactions from both the pro and con camps, and the appearance of a good-looking, fast electric car stoked everyone’s passions. Tesla had also turned Silicon Valley into a real threat, at least conceptually, to Detroit for the first time. The month after the Santa Monica event was the Pebble Beach Concours d’Elegance, a famous showcase for exotic cars. Tesla had become such a topic of conversation that the organizers of the event begged to have a Roadster and waived the usual display fees. Tesla set up a booth, and people showed up by the dozens writing $100,000 checks on the spot to pre-order their cars. “This was long before Kickstarter, and we just had not thought of trying to do that,” Tarpenning said. “But then we started getting millions of dollars at these types of events.” Venture capitalists, celebrities, and friends of Tesla employees began trying to buy their way onto the waiting list. Some of Silicon Valley’s wealthy elite went so far as to show up at the Tesla office and knock on the door, looking to buy a car. The entrepreneurs Konstantin Othmer and Bruce Leak, who had known Musk from his internship days at Rocket Science Games, did just that one weekday and ended up getting a personal tour of the car from Musk and Eberhard that stretched over a couple of hours. “At the end we said, ‘We’ll take one,’” Othmer said. “They weren’t actually allowed to sell cars yet, though, so we joined their club. It cost one hundred thousand dollars, but one of the benefits of membership was that you’d get a free car.”
As Tesla switched from marketing back into R&D mode, it had some trends working in its favor. Advances in computing had made it so that small car companies could sometimes punch at the same weight as the giants of the industry. Years ago, automakers would have needed to make a fleet of cars for crash testing. Tesla could not afford to do that, and it didn’t have to. The third Roadster engineering prototype went to the same collision testing facility used by large automakers, giving Tesla access to top-of-the-line high-speed cameras and other imaging technology. Thousands of other tests, though, were done by a third party that specialized in computer simulations and saved Tesla from building a fleet of crash vehicles. Tesla also had equal access to the big guys’ durability tracks made out of cobblestones and concrete embedded with metal objects. It could replicate 100,000 miles and ten years of wear at these facilities.
Quite often, the Tesla engineers brought their Silicon Valley attitude to the automakers’ traditional stomping grounds. There’s a break and traction testing track in northern Sweden near the Arctic Circle where cars get tuned on large plains of ice. It would be standard to run the car for three days or so, get the data, and return to company headquarters for many weeks of meetings about how to adjust the car. The whole process of tuning a car can take the entire winter. Tesla, by contrast, sent its engineers along with the Roadsters being tested and had them analyze the data on the spot. When something needed to be tweaked, the engineers would rewrite some code and send the car back on the ice. “BMW would need to have a confab between three or four companies that would all blame each other for the problem,” Tarpenning said. “We just fixed it ourselves.” Another testing procedure required that the Roadsters go into a special cooling chamber to check how they would respond to frigid temperatures. Not wanting to pay the exorbitant costs to use one of these chambers, the Tesla engineers opted to rent an ice cream delivery truck with a large refrigerated trailer. Someone would drive a Roadster into the truck, and the engineers would don parkas and work on the car.
Every time Tesla interacted with Detroit it received a reminder of how the once-great city had been separated from its own can-do culture. Tesla tried to lease a small office in Detroit. The costs were incredibly low compared with space in Silicon Valley, but the city’s bureaucracy made getting just a basic office an ordeal. The building’s owner wanted to see seven years of audited financials from Tesla, which was still a private company. Then the building owner wanted two years’ worth of advanced rent. Tesla had about $50 million in the bank and could have bought the building outright. “In Silicon Valley, you say you’re backed by a venture capitalist, and that’s the end of the negotiation,” Tarpenning said. “But everything was like that in Detroit. We’d get FedEx boxes, and they couldn’t even decide who should sign for the package.”
Throughout these early years, the engineers credited Eberhard with making quick, crisp decisions. Rarely did Tesla get hung up overanalyzing a situation. The company would pick a plan of attack, and when it failed at something, it failed fast and then tried a new approach. It was many of the changes that Musk wanted that started to delay the Roadster. Musk kept pushing for the car to be more comfortable, asking for alterations to the seats and the doors. He made the carbon-fiber body a priority, and he pushed for electronic sensors on the doors so that the Roadster could be unlocked with the touch of a finger instead of a tug on a handle. Eberhard groused that these features were slowing the company down, and many of the engineers agreed. “It felt at times like Elon was this unreasonably demanding overarching force,” said Berdichevsky. “The company as a whole was sympathetic to Martin because he was there all the time, and we all felt the car should ship sooner.”
By the middle of 2007, Tesla had grown to 260 employees and seemed to be pulling off the impossible. It had produced the fastest, most beautiful electric car the world had ever seen almost from thin air. All it had to do next was build a lot of the cars—a process that would end up almost bankrupting the company.
The greatest mistake Tesla’s executives made in the early days were assumptions around the transmission system for the Roadster. The goal had always been to get from zero to 60 mph as quickly as possible in the hopes that the raw speed of the Roadster would attract a lot of attention and make it fun to drive. To do this, Tesla’s engineers had decided on a two-speed transmission, which is the underlying mechanism in the car for transferring power from the motor to the wheels. The first gear would take the car from zero to 60 mph in less than four seconds, and then the second gear would take the car up to 130 mph. Tesla had hired Xtrac, a British company sp
ecializing in transmission designs, to build this part and had every reason to believe that this would be one of the smoother bits of the Roadster’s journey. “People had been making transmissions since Robert Fulton built the steam engine,” said Bill Currie,8 a veteran Silicon Valley engineer and employee No. 86 at Tesla. “We thought you would just order one. But the first one we had lasted forty seconds.” The initial transmission could not handle the big jump from the first to the second gear, and the fear was that the second gear would engage at high speed and not be synchronized with the motor properly, which would result in catastrophic damage to the car.
Lyons and the other engineers quickly set out to try to fix the issue. They found a couple of other contractors to design replacements and again hoped that these longtime transmission experts would deliver something usable with relative ease. It soon became apparent, however, that the contractors were not always putting their A team to work on this project for a tiny start-up in Silicon Valley and that the new transmissions were no better than the first. During tests, Tesla found that the transmissions would sometimes break after 150 miles and that the mean time between failures was about 2,000 miles. When a team from Detroit ran a root cause analysis of the transmission to find failures, they discovered fourteen separate issues that could cause the system to break. Tesla had wanted to deliver the Roadster in November 2007, but the transmission issues lingered, and by the time January 1, 2008, rolled around, the company had to once again start from scratch, on a third transmission push.
