The electric car manufacturer says updates to its battery design and manufacturing process will help lower production costs.
- The high cost of batteries is the main reason why electric vehicles cost more than gas-powered cars.
- At the company's 'Battery Day' event on Tuesday, Tesla announced a new battery design that will give its cars more power and a longer range.
- The success of Tesla's plan depends on its ability to scale up production.
Cheaper, more efficient batteries. That's what Tesla says will allow it to offer a $25,000 electric car within the next three years. The announcements came at the company's "Battery Day" event on Tuesday afternoon in Palo Alto, California.
"One of the things that troubles me the most is that we don't yet have a truly affordable car, and that is something that we will make in the future," Tesla CEO Elon Musk told a socially distanced audience, who were sitting in cars in a parking lot. "But in order to do that, we've got to get the cost of batteries down."
How to cut costs? Tesla is working on a design update for its batteries, and the company plans to begin manufacturing them in-house. (Panasonic currently produces Tesla batteries.) A key design update is removing a tab within the battery that connects the cell to what it powers.
"You actually have a shorter path length [for the electron to travel] in a large tabless cell than you have in the smaller cell with tabs," Musk said. "So even though the cell is bigger, it actually has more power."
Tesla also plans to lower costs by using nickel instead of cobalt in its cathodes. The company said its new cathode design would reduce costs by about 75 percent, and also remove waste water from the manufacturing process.
What's more, the international cobalt supply is limited, and most of it comes from the Democratic Republic of Congo, where adult and child miners are known to be exploited.
Screenshot of Tesla's 'Battery Day' presentation
It's unclear when Tesla will stop using cobalt, or when it will stop sourcing its batteries from Panasonic. But Tesla claims that its new battery design and manufacturing changes will allow it to cut the cost per kilowatt-hour in half. If Tesla can successfully scale up production, the company could hit its goal of $100 per kilowatt-hour sooner than expected.
Hitting that mark could usher in the electric-car revolution, considering $100 per kilowatt-hour is generally regarded as the threshold the industry needs to reach in order to make electric vehicles cost competitive with gas-powered cars.
A $25,000 electric car would also be Tesla's cheapest offering by far. The company had previously promised a $35,000 car, but only offered one at that price for a limited time. Tesla's website says its Model 3, its cheapest car, starts at about $39,000.
Photo of Tesla's new battery design
To be sure, Musk is known for promising big on his projects, but not always following through on the promised timetable. But despite having an "insanely hard" 2020, as Musk said, Tesla's had a good past couple years.
"In 2019, we had 50% growth," Musk said at the event. "And I think we'll do really pretty well in 2020, probably somewhere between 30 to 40 percent growth, despite a lot of very difficult circumstances."
So far, 30 student teams have entered the Indy Autonomous Challenge, scheduled for October 2021.
- The Indy Autonomous Challenge will task student teams with developing self-driving software for race cars.
- The competition requires cars to complete 20 laps within 25 minutes, meaning cars would need to average about 110 mph.
- The organizers say they hope to advance the field of driverless cars and "inspire the next generation of STEM talent."
The Indianapolis Motor Speedway has hosted thousands of high-speed races since it was constructed in 1909. But none were like the upcoming Indy Autonomous Challenge, in which student teams will develop software for self-driving cars and race them for a chance to win $1.5 million in prizes.
First announced in 2019, the competition is scheduled for October 23, 2021. It's set to be a head-to-head, 20-lap race, with each team having 25 minutes to complete those laps.
So far, 30 student teams from universities across four continents have entered the competition. Each team will be given the same race car — a modified Dallara IL-15 — and a set of hardware, including sensors and computers.
But students must develop their own self-driving software, including "neural nets, computer vision and other artificial intelligence systems that will allow the cars to race at high speeds," according to The Wall Street Journal.
Indy Autonomous Challenge
Completing the race in 25 minutes means the cars will need to average about 110 miles per hour. So, while the race may end up being a bit slower than a typical Indy 500 competition, in which winners average speeds of over 160 mph, it's still set to be the fastest autonomous race featuring full-size cars.
"There is no human redundancy there," Matt Peak, managing director for Energy Systems Network, a nonprofit that develops technology for the automation and energy sectors, told the Pittsburgh Post-Gazette. "Either your car makes this happen or smash into the wall you go."
Illustration of the Indy Autonomous Challenge
Indy Autonomous Challenge
The Indy Autonomous Challenge describes itself as a "past-the-post" competition, which "refers to a binary, objective, measurable performance rather than a subjective evaluation, judgement, or recognition."
This competition design was inspired by the 2004 DARPA Grand Challenge, which tasked teams with developing driverless cars and sending them along a 150-mile route in Southern California for a chance to win $1 million. But that prize went unclaimed, because within a few hours after starting, all the vehicles had suffered some kind of critical failure.
Indianapolis Motor Speedway
Indy Autonomous Challenge
One factor that could prevent a similar outcome in the upcoming race is the ability to test-run cars on a virtual racetrack. The simulation software company Ansys Inc. has already developed a model of the Indianapolis Motor Speedway on which teams will test their algorithms as part of a series of qualifying rounds.
"We can create, with physics, multiple real-life scenarios that are reflective of the real world," Ansys President Ajei Gopal told The Wall Street Journal. "We can use that to train the AI, so it starts to come up to speed."
Still, the race could reveal that self-driving cars aren't quite ready to race at speeds of over 110 mph. After all, regular self-driving cars already face enough logistical and technical roadblocks, including crumbling infrastructure, communication issues and the fateful moral decisions driverless cars will have to make in split seconds.But the Indy Autonomous Challenge says its main goal is to advance the industry, by challenging "students around the world to imagine, invent, and prove a new generation of automated vehicle (AV) software and inspire the next generation of STEM talent."
Help future Mars rovers better navigate the red planet's treacherous terrain.
- NASA just announced its AI4Mars project, which lets you can take a virtually simulated tour around Mars via the Curiosity rover.
- The simulation project is calling on users to help the rover better classify the planet's sometimes dangerous terrain by labeling images taken by Curiosity.
- This project gives you a chance to participate in enhancing the new machine learning approaches for exploring Mars and unveiling its secrets.
If you've ever wanted a close-up experience of what it's like to roam on Mars' surface, now is your chance. NASA just announced its AI4Mars project, which lets you can take a virtually simulated tour around the Red Planet via the Curiosity rover.
Improving future rovers
Designed by a team at NASA's Jet Propulsion Laboratory, the simulation project is calling on users to help the rover better classify the planet's sometimes dangerous terrain. Your task is to identify and label images taken by Curiosity from Mars' surface for scientists to use. The crowdsourced data will help train a future rover to more safely navigate obstacles like bedrocks or sand.
Mars rovers have an unfortunate habit of getting stuck in sand traps, and sometimes never getting out, as was the tragic fate of NASA's Spirit Rover. The project hopes to make future rovers similar to self-driving vehicles that know "where it's safe to drive, land, sleep and hibernate," according to the website.
How it works
When you open the classification tool on the website, you're instructed to select different surface types — sand, soil, bedrock, and big rocks — using a polygon drawing tool designated for the type of terrain you are labeling. After you've identified everything in the image, you click "Done" to move on to the next photo and do it again. If you aren't sure about an object, the website asks you to leave it unlabeled. It also asks you not to overlap the polygons. If you get confused, click "Tutorial" to open a popover and a discussion board where you can ask questions.
You won't get to virtually control Curiosity around the surface of Mars like a video game. But this project does give you a chance to get an intimate look at the planet's surface and enhance the new machine learning approaches for exploring mars and unveiling its secrets.
Similar projects calling on volunteers to help with scientific research can be found at Zooniverse's project page. For example, you can help researchers find asteroids in images from the Hubble Space Telescope, or help Seismologists by listening for Earthquakes using technology that makes seismic waves audible.
Such a battery would make it far cheaper to implement robotaxis and long-haul electric trucks, both of which Tesla is developing.
- A team of researchers working with Tesla recently released a paper describing a lithium-ion battery that should last 1 million miles over 4,000 charges and depletions.
- The researchers reportedly optimized commonly used components of EV batteries, and made their findings available to other battery researchers.
- Tesla CEO Elon Musk said robotaxis could hit streets as early as 2020.
Tesla seems to have made good on CEO Elon Musk's promise from earlier this year to develop an electric-vehicle battery with a lifespan of more than 1 million miles, according to a recent paper and patent. A million-mile battery would roughly double the lifetime of batteries currently used in Tesla cars, and also significantly cut the operating costs of robotaxis and long-haul electric trucks, both of which Tesla is developing.
Tesla has an exclusive agreement with a group of battery researchers — led by Jeff Dahn, a physics professor who some call Tesla's "battery guru" — and earlier this month they published a paper describing a lithium-ion battery with a longer lifespan and significantly higher energy capacity than what's currently on the market.
The battery should last more than 1 million miles over 4,000 charges and depletions, all while losing less than 10 percent of its energy capacity over its lifetime, according to the paper. That's a significant upgrade compared to lithium-ion batteries from about five years ago that lost half of their energy capacity after 1,000 charge cycles, as Wired notes.
"4,000 cycles is really impressive," Greg Less, the technical director at the University of Michigan's Energy Institute battery lab, told Wired. "A million mile range is easily doable with 4,000 cycles."
The researchers provided specific details about the battery in their paper to help advance research in the field.
"Full details of these cells including electrode compositions, electrode loadings, electrolyte compositions, additives used, etc. have been provided," Dahn and his colleagues wrote in the paper. "This has been done so that others can recreate these cells and use them as benchmarks for their own R+D efforts."
The new battery doesn't bring anything especially new to the table in terms of chemical composition. Rather, Dahn and his colleagues managed to optimize commonly used components of EV batteries — lithium nickel manganese cobalt oxide, or NMC. The battery uses relatively large NMC crystals in the cathode (a battery's positive electrode), which helps the battery last longer by preventing it from cracking during charging.
Tesla — the company's cars already go farther on a single charge than other manufacturers — likely isn't losing its competitive edge by sharing these details: Just days after the recent paper was published, Tesla and Dahn received a patent for a single-crystal lithium-ion battery similar to the one described in the paper, but with an electrolyte additive that'll likely enable it to perform even better, enhancing the "performance and lifetime of Li-ion batteries, while reducing costs," as the patent states.
Such a battery would help Tesla implement robotaxis and long-haul electric trucks at lower costs. But that doesn't necessarily mean consumers should expect the cost of Teslas to drop anytime soon. In fact, Musk said in July that the price of Teslas will likely surge once the company's robotaxis hit the streets, considering that supply and demand will make the cars even more valuable.
@DisruptResearch @flcnhvy @lexfridman @Tesla To be clear, consumers will still be able to buy a Tesla, but the clea… https://t.co/D0k2MEVogE— Elon Musk (@Elon Musk)1562564988.0
The 2020 Democratic candidate's plan to give Americans a universal basic income seems to include a special provision for truckers.
Photo credit: ROSLAN RAHMAN / Contributor
- Andrew Yang is a 2020 Democratic presidential candidate who's made technology and automation central to his campaign.
- Yang says the U.S. needs a plan for how to manage the upcoming loss of millions of American trucking jobs to self-driving vehicles.
- Yang wants to tax profits from self-driving trucks to give these laid-off truckers a "severance package."
At the core of Andrew Yang's 2020 presidential campaign is an existentially unsettling message: Automation is coming for our jobs and it's going to restructure the economy. In fact, it already is, according to the candidate.
"Technology is now automating away millions of American jobs," he said during a Democratic primary debate in June. "It's why Donald Trump is our president today — that we automated away 4 million manufacturing jobs in Michigan, Ohio, Pennsylvania, and Wisconsin, and we're about to do the same thing to millions of retail jobs, call center jobs, fast food jobs, truck driving jobs and on and on through the economy."
The success of Yang's campaign — which, to be sure, is relatively minor, considering he's polling at about 3 percent as of August 22 — can be attributed in part to the fact that he's offering a potential way out of this mess: give every American adult a universal basic income of $1,000 per month. That plan, as the Democratic hopeful recently described on his website, needs to give special consideration to the millions of Americans whose jobs will likely be automated sooner than others: truckers.
There are 3.5 million truck drivers in America. Another 5 million work in the truck stops, motels, and diners that… https://t.co/nhqfalyKge— Andrew Yang (@Andrew Yang)1551383445.0
A statement on Yang's campaign website reads:
"Over 3 million Americans work as truck drivers, and over 7 million are employed related to trucking activity. Self-driving truck technology is rapidly becoming sophisticated enough to replace these drivers, and the economy is not prepared to absorb the loss of so many jobs. Truck drivers are 94% male, average age 49, average education high school or one year of college – there are not necessarily other opportunities for them that will pay a comparable salary. Additionally, hundreds of communities are built around the trucking industry and those communities are also at risk from the coming automation."
How can the U.S. "ease the transition" to self-driving vehicles, as Yang's website describes? Tax profits earned from self-driving trucks to provide a severance package for out-of-work truckers.
"The estimated cost-savings and efficiency gains of automated freight are $168 billion per year which is enough to pay the truckers significant sums and still save tens of billions per year," Yang's website states.
Automated trucks: Blue-collar disaster or economic win?
Of course, what's unclear is how accurate those estimates are, and how exactly Yang would go about taxing the self-driving trucking industry (though we do know who would get the ball rolling – the so-called "Trucking Czar" Yang would appoint if elected president). What's more, Yang – the only candidate who's made tech and automation central to his campaign — could be wrong about how imminent of a threat automation is to the economy. But recent developments in the industry seem to suggest it is, in fact, a looming problem.
American companies are already experimenting with self-driving trucks. In 2019, the United States Postal Service, UPS and Amazon worked with the self-driving trucking company TuSimple to run pilot programs that involved shipping cargo on self-driving trucks. In these test runs, the self-driving trucks operated at "Level 4" autonomy, as measured by the Society of Automotive Engineers' "Levels of Driving Automation" — this means that the trucks drove automatically but there were, in this case, two people inside the cabin at all times, ready to take the wheel in the event of an emergency.
Before self-driving trucks can hit the roads in large numbers, they'll need to pass a set of regulatory hurdles, and it's unclear how long that would take. But on the technology side, the trucks could reach full autonomy by the end of 2020, according to TuSimple President Xiaodi Hou.