On July 26, in a rare glimpse into progress on his company's purported game-changer battery, Tesla CEO Elon Musk was in true Muskian form, raising hopes and tempering expectations in one cryptic swoop:
We have successfully validated performance and lifetime of our 4680 cells produced at our Kato facility in California. We are nearing the end of manufacturing validation at Kato: field quality and yield are at viable levels and our focus is now on improving the 10% of manufacturing processes that currently bottleneck production output. While substantial progress has been made, we still have work ahead of us before we can achieve volume production. Internal crash testing of our structural pack architecture with a single-piece front casting has been successful.
Any whisper of production bottlenecks at Tesla inevitably recalls the Model 3's troubled launch in 2017 and 2018, the one that had Musk sleeping in his office. Yet that make-or-break sedan ultimately made Tesla the world leader in EVs, and one of its most-valued companies.
The enlarged, cylindrical 4680 cell, which Tesla first teased at its Battery Day last September, brings its own sky-high hopes and challenges: If Tesla can pull off in-house, vertically integrated battery manufacturing, and the cell performs as advertised, the 4680 could fuel Musk's dreams to build millions of EVs a year around the world. Tesla's goals include boosting driving range by more than 50 percent—16 percent of that due to the 4680's newfound punch—while halving battery costs and bringing a $25,000 Tesla to showrooms. Tesla continues to dominate EV sales in America, but its seemingly insurmountable lead in driving range is under assault. The Arizona-built Lucid Air sedan, the work of Musk's former Model S chief engineer, has demonstrated it can travel up to 517 miles, a lofty record for any EV. Tesla's best, the Model S Long Range, is EPA-rated for 405 miles.
"The Lucid Air is the first car to show range that's not just competitive (with Tesla), but better, an astonishing achievement," said Venkat Viswanathan, battery researcher and associate professor of mechanical engineering at Carnegie Mellon University. "It shows it's no longer a one-horse race."
To gallop back in front, a 50-percent range leap for a vehicle like the Model S would let it top 600 miles, a diesel-like stamina that seemed unimaginable a few years ago. So much for "range anxiety." Sandy Munro, the Detroit-area engineer who has gained YouTube fame for his reverse-engineered teardowns and analyses of EVs, is among the experts convinced that Tesla will pull it off.
"For the cell itself, no question, it will kick the daylights out of everybody," Munro says.
The 4680 could fuel Musk's dreams to build millions of EVs a year around the world
That kick begins with the 4680's form factor and what surrounds it, more than what ends up inside it, Munro and other experts told IEEE Spectrum. Where Battery Day trumpeted the new cell as having five times the energy density and six times the power of its Panasonic-built 2170 cells, Musk conveniently failed to mention the larger 4680 has nearly 5.5 times the volume, simply due to larger dimensions: 46-by-80 millimeters, versus 21-by-70 millimeters. Yet this "bigger can" brings big benefits. Each jelly-roll cell packs in more active battery material, and less wasted space in metal casing. A so-called "structural battery pack" (also called "cell-to-pack" construction), touted as a Tesla innovation, is in fact already a staple of several EVs, especially in China—including General Motors' red-hot, roughly $5,000 Wuling Mini. That saves more space by trading multiple module cases for a more-streamlined pallet of cells wired in parallel.
"It's basically a giant box without dividers," Munro says.
And where current Tesla packs feature what Munro calls "a crappy cooling design" with thermal channels between cells—superfluous, because most heat is concentrated at cell ends—Munro says 4680 cells will rest atop a liquid-cooled thermal plate that's become a staple of EVs from GM, Ford, Volkswagen, Porsche, or Rivian.
"When we first tore apart the Model 3, we just couldn't figure that out," Munro says of the previous method.
Munro further estimates Tesla's redesigned pack, including adhesive bonds between cells and modern welding techniques, will reduce steel use by 30 to 40 percent. Stamped grid plates on top will bring power back to terminals. Munro's team mocked up the projected pack, including cut and painted wooden dowels to mimic the beefy new cell. That pack swallowed 960 larger cells, versus 4,416 cells for the 2170 variety. Totaling potential gains, Munro estimates Tesla could stuff 130 kWh of new cells into the same-sized pack that houses just 72 kWh in the Model 3.
His analysis suggests a 4680 cell with roughly 9,000 mAH, versus 5,000 mAH for the 2170. Munro cautions these aren't definitive estimates; Tesla has yet to show its 4680 cell in physical form, or reveal its chemistry or specs. Still, experts say a long list of innovations could widen Tesla's already significant lead in driving range and efficiency versus the global giants.
"The holistic approach to EV and battery engineering is Tesla's key advantage," Munro says.
That chain-of-gains approach includes a "tabless" cell design, which some experts see as the 4680's best physical innovation. Eliminating traditional tabs that connect a cathode and anode to battery terminals simplifies manufacture, saves space and reduces ohmic resistance, a major hurdle in safely charging a large-format battery.
Elon Musk shared designs for the 4680 at Battery Day 2020.Tesla
"You actually have a shorter path length in a large tabless cell than you have in a smaller cell with tabs," Musk explained, a year ago at Tesla's Battery Day.
Musk actually cites Tesla's growing manufacturing expertise in batteries—eliminating steps, streamlining processes, slashing costs—as its true competitive edge. That vision includes not just 4680 factories adjacent to car plants near Austin and Berlin but also chemical plants to produce cathodes and lithium hydroxide, according to Simon Moores, managing director of Benchmark Mineral Intelligence. In one example, Tesla plans to use raw metallurgical silicon to boost its content in cells, using a scalable elastic polymer coating to conduct ions, boosting range another 20 percent and reducing pack costs by five percent, according to Drew Baglino, its SVP of engineering.
Two weeks before Battery Day, Tesla bought three patent applications for just $3 from Springpower International, an obscure Canadian company that Tesla now seems to have acquired outright. One Springpower invention, described by Musk and Baglino at Battery Day, uses recirculation to skip the step of treating contaminated water in cathode production—up to 4,000 gallons of effluent for every ton of cathode material. The same process might ease battery recycling and grid-storage solutions.
"It's insanely complicated, like digging a ditch, filling it in and digging the ditch again," Musk said. "We looked at the entire value chain and said, 'How can we make this as simple as possible'"?
Meanwhile, EV demand threatens to outstrip battery supply, kneecapping EV production and adoption, and delaying a climate-critical switch from fossil fuels to electricity
An electrode-coating process using dry film, linked to another Tesla supplier setting up shop in Texas, would eliminate toxic, expensive solvents used in aqueous coatings.
It all sounds amazing, on paper. Gary Koenig, a battery materials expert and associate professor of chemical engineering at the University of Virginia, cautioned that getting the cells into mass production at cost is another story.
"Gettings these things to be scalable is really hard," Koenig says of Tesla's 10-gigawatt pilot plant. "And you have to get to high volume to bring those costs down, and that's not easy to do."
Viswanathan agrees that Battery Day blended tangible gains with Musk's familiar visionary spitballing, making it all hard to parse.
"It's difficult to separate the signal from the noise in these Tesla presentations," Viswanathan says. "But there's always signal; the question is how much."
"Still, I have no doubt they will be able to produce the cell at volume," he continued. "More than any automaker, they have the talent to do it. The question really is how quickly, at what price point, safety levels and manufacturing defects."
As for chemistry, Viswwanathan said a nickel-rich, low-cobalt design will surely power pricier, longer-range models, to take on rivals including Lucid and General Motors—the latter with its new nickel-intensive Ultium cells. Regarding Tesla's vaguely stated "diversified cathode" strategy, Viswanathan said it's possible that Tesla would build some lithium-iron phosphate (LFP) batteries in the 4680 format. That chemistry, once dismissed for its puny energy density, is suddenly prized for low cost, long life and reassuring safety; especially for entry-priced models. Tesla has already begun selling Model 3s in China and Europe with CATL's prismatic LFP cells. And on August 27, Tesla dangled LFP-powered versions of the Model 3 Standard Range Plus to Americans awaiting backlogged deliveries. Tesla contacted some reservation holders, offering them a car as early as September if they agreed to accept a Model 3 SR+ with an LFP pack with an estimated 253-mile range, 10 fewer miles than versions with 2170 batteries.
Musk himself stated that LFP may make up 50 percent of all lithium-ion cells in cars, versus fewer than 10 percent today. Viswanathan notes that, by packing in more active battery material via 4680 cells and structural packs, LFP could enjoy a serious boost in driving range at an ultra-affordable cost.
So, when might the world actually drive Teslas powered by the 4680? The question takes on more urgency, considering Tesla's frustrating record of tardiness. In late April, Musk said the battery was 12 months away from production, if not 18 months. Existing suppliers, including Panasonic, CATL, LG Energy Solution and SK Innovation, may well deliver the 4680 before Tesla itself. (Panasonic's new chief executive confirmed his company will make a large investment to build 4680 cells if they prove viable).
After months of playing coy, Tesla finally confirmed in August that the Cybertruck's Texas rollout would be pushed back to 2022, due to battery shortages. That massively scaled pickup is one power-hungry candidate for the 4680 cell, along with the also-delayed commercial Semi.
Cybertruck, unveiled to great fanfare in 2019, will now be beaten to market by this year's GMC Hummer pickup, the Rivian R1T and perhaps even Ford's F-150 Lightning in 2022. Model Y SUV's from Giga Texas and Giga Berlin are also in line for the 4680, built at those factories' adjacent battery-production lines. Tesla continues to hedge its bets, expanding existing contracts with Panasonic and other suppliers.
Tesla's Battery Day targets, meanwhile, include a seemingly quixotic goal to ramp up battery production (including from supplier partners) to 3 terawatts by 2030. That's a nearly 100-fold jump from Tesla's current capacity in Nevada. It would be enough to supply 20 million annual Teslas (or rival automakers who buy Tesla batteries), versus roughly 500,000 global Tesla sales in 2020.
Yet that ambition is driven by harsh reality: A global tsunami of EV demand threatens to outstrip battery supply, kneecapping EV production and adoption, and delaying a climate-critical switch from fossil fuels to electricity. Energy research company Wood Mackenzie estimates EVs will grow from 10 million on roads today, to 100 million in 2030 and 400 million after 2040. EVs would hog 90 percent of battery demand over next two decades. Even at Wood Mackenzie's conservative estimate, battery demand would be about eight times as much as today's factories can deliver. To Tesla, it's all about a Marshall-Plan level campaign to build not just battery factories, but factories that churn out more batteries, more affordably—for its own global domination, and to achieve Musk's more-altruistic goals.
"They've been constrained for years on batteries," Viswanathan says. "They see this as a critical piece of their growth story, to meet the volume demand that they and the industry will have in coming years."