Musk Sets Tesla Battery Deadline After Setbacks 

Two of those people said they weren’t told what the consequences would be if the team misses the end of year deadline. However, both believe Musk may throw in the towel on trying to scale up the 4680 battery in house and instead rely largely on outside suppliers for the technology. 

Such a change would represent a major reversal on a technology Musk previously described as so critical to Tesla’s future that it had to make its own. His motivations for the battery plan echoed Apple’s decision more than 15 years ago to begin designing its own processors for the iPhone. In Apple’s case, the move was a stunning success, allowing it to deliver speed and battery performance that rivals have a hard time matching with off-the-shelf processors. 

Recently, Musk made a similar decision to go in-house with another one of his companies, xAI. The Information reported last week that xAI has abandoned talks about using Oracle’s cloud service to run its artificial intelligence models and plans to build its own data centers instead. 

“Oracle is a great company…but, when our fate depends on being the fastest by far, we must have our own hands on the steering wheel, rather than be a backseat driver,” Musk said in a post on X, confirming the news. 

In the case of Tesla, the passion Musk once had for using batteries to differentiate the company’s products seems to have given way to a fervor for another technology: AI. Earlier this year, he promised to unveil the robotaxi—a steering wheel–free vehicle that will be powered by Tesla’s Full Self-Driving software—at an event on Aug. 8 (Bloomberg reported last week that the event has been postponed until October). 

And he has raved lately about the opportunity represented by Optimus, a humanoid robot that he claimed in June could one day give Tesla a $25 trillion valuation. 

Perfecting the AI in those products could prove as devilish as the technical challenges Tesla has faced with its battery ambitions. Tesla already has succeeded in making a version of the 4680, though without one of the key innovations that Musk promised would set it apart from more conventional batteries. The more conventional version has underperformed in its initial use in Tesla’s Model Y and Cybertruck (Tesla has stopped using it in the Model Y). 

Rather than the large boost in energy density and range that Musk predicted, that conventional version of the 4680 has so far performed on par with middle-of-the-road conventional batteries, according to Jordan Giesige, a Tesla analyst who runs the podcast “The Limiting Factor.” 

At the same time, prices on batteries from outside suppliers have plummeted, making it more attractive to buy them from companies like South Korea’s LG Energy Solution and Japan’s Panasonic. Musk’s rhetoric on its in-house batteries has shifted accordingly.

Rather than describing them as a key to differentiating Tesla’s cars, he has wholly revised his rationale for developing the battery. In April, Musk and his senior aides told analysts that the 4680 had simply been a “hedge” against runaway battery prices that had gone “bonkers.”

Battery Day

Musk’s retreat from his ambitious battery plans comes at a challenging time for Tesla, which this year may post its first annual sales decline since 2010. In April, Musk let go of more than 10% of Tesla’s global workforce. Last week, a market research firm said Tesla’s share of U.S. EV sales fell below 50% in the second quarter for the first time.

Tesla’s outlook was very different at the September 2020 event in front of Tesla investors and owners, which Musk called Battery Day. At the time, the company had just posted its first-ever string of four consecutive profitable quarters, prompting other automakers to accelerate their own EV plans. Musk wanted to demonstrate how Tesla planned to leave its rivals even further behind.

On a stage in the parking lot of Tesla’s Fremont, Calif., assembly factory, Musk and Drew Baglino, then Tesla’s senior vice president for powertrain, laid out a road map for Tesla and other companies to replace 1 billion gasoline cars—every combustion vehicle on the planet—with EVs.

For Tesla, that meant producing and selling 20 million EVs annually, a roughly 40-fold increase from the company’s sales that year. To sell that many vehicles, Tesla would need to make an EV almost anyone could afford, something Musk said had been his “dream from the beginning of the company.” What’s more, the car also had to be “very compelling,” Musk said, so that not only could almost anyone afford it, but almost anyone would want it.

To make that $25,000 car, Musk and Baglino said they needed to slash the cost of the battery, typically the most expensive component in an EV, accounting for as much as 40% of the vehicle’s cost. To do that, Tesla needed to take charge of the whole enterprise, “from mining the ore to a complete battery pack,” Musk said.

They displayed slides showing how Tesla would develop processes for nickel and lithium to be mined and refined more cheaply and cleanly. They said their new 4680 battery—twice the size of the next largest cylindrical EV battery, the 2170, which powers Tesla’s Y and 3 models—would pack five times the energy and six times the power of the best available conventional batteries, and would also boost driving range 16%.

To produce that battery, Musk and Baglino said they needed to scrap the usual way of making the electrodes—the anode and cathode, which serve as the catcher’s mitts for the lithium ions shuttling back and forth in a battery, creating the energy that propels an EV.

The production of electrodes is among the dirtiest, costliest parts of the whole affair. The traditional process of creating cathodes required the use of a toxic solvent, 100-meter-long ovens, enormous quantities of energy and a factory the size of an aircraft hangar.

By eliminating this “wet” process, as the industry calls it, Tesla could cut the battery’s overall cost 15%, Musk said; other industry hands estimated that the savings could reach 30%. The company planned to replace that method with a more novel “dry” process of making electrodes, aided by Tesla’s 2019 acquisition of a San Diego startup, Maxwell Technologies, that had pioneered the technique.

The dry process relied on smaller equipment, and required no toxic solvent and no ovens. Tesla would need a facility only a tenth of the size of the hangarlike space required for the wet production process.

By Tesla’s calculation, if it incorporated dry electrodes and added a slew of other innovations Musk also described on Battery Day, it would reduce the battery’s cost 56%, bringing it below $60 per kilowatt-hour. That would be well under the $100 per kilowatt-hour threshold at which EVs begin to cost less than similar combustion models.

“We’re not getting into the cell business just for the hell of it,” Musk said at the Battery Day event. “It’s because it’s the fundamental constraint; it’s the thing that is the limiting factor for rapid [EV sales] growth.”

The announcement jolted the EV industry. Soon after, battery makers Panasonic, LG, China’s BAK Battery and Eve Energy, and a slew of other companies all said they, too, would develop versions of the 4680. Volkswagen, Siemens, and Japanese battery maker Asahi Kasei said they would work on dry electrode processing.

Hieu Duong, a Tesla employee who had been an architect of dry electrode processing at Maxwell, left the company to work on the process for a startup called AM Batteries.

Still, Musk struck one cautionary note about Tesla’s battery plans: The process for making dry electrodes “does work,” he said, “but with not a good, not a high yield.”

Pizza Dough Hell

Ten months later, the challenges Musk alluded to became more clear.

Asked about Tesla’s progress during a quarterly earnings call in July 2021, he noted that the company was having trouble with the dry production process. Workers were finding that the dry powder was seriously damaging production equipment.

In the usual production method, workers mix graphite, nickel, lithium and cobalt powders with a solvent or water to make a slurry, which they coat onto giant sheets of foil. They then feed the coated foil sheets along a conveyor belt through huge rollers called calenders, which flatten the mixture uniformly. The foil sheets then become the two electrodes in a battery.

In the dry process, workers flatten the powders without any solvent or water. On the anode side of the battery, that’s relatively easy because it involves a single, soft and relatively malleable material—graphite.

But making the cathodes can be tricky, because they are composed of nickel, cobalt and lithium, whose particles can easily clump. One battery company CEO equated making a dry cathode to baking a cake without an egg: It can be done, but you can also end up with a pile of crumbs.

Musk used a similar analogy when describing the adversity Tesla was having with the new production method. Tesla workers were finding the dry experience like “squashing…very hard pizza dough. And it’s denting the calender rollers,” he said. Musk laughed, but he was clearly frustrated.

He also began to walk back his statements at Battery Day about how core dry cathodes would be to Tesla’s success. The dry method would help cut costs, he said, “but it’s not the Messiah.”

Still, Tesla seemed to be making progress. In November 2021, Musk drove a Cybertruck onto a stage before a cheering crowd at Giga Texas, the company’s Austin, Texas, factory. It was the first Tesla powered by the 4680 battery. And in early 2023, Tesla’s Baglino said on an earnings call that the company was producing enough batteries to make 1,000 vehicles a week.

But Tesla was cutting corners to get the 4680s it was producing out the door: In the first independent teardown of the 4680, battery researchers at the University of California, San Diego found that Tesla was dry-coating only the anode. It was creating the cathodes using the more conventional wet process, purchased from outside Tesla’s suppliers, according to people with knowledge of the matter.

Other researchers found the same thing. In a July 2023 teardown of the battery, Japanese battery analyst Hideo Takeshita said the 4680 was no more energy dense than conventional batteries, meaning it would not deliver greater range. “All the hard work that went into making them has gained no performance advantage,” Takeshita wrote.

Talking to analysts that month, Baglino conceded as much. “We’re yet to integrate silicon or in-house cathode production, both reviewed on Battery Day, which do bring significant further energy density and cost improvements, but that is a topic for another day,” he said.

An Ominous Exit

This year, tumult inside Tesla has raised further doubts about how important battery innovation is to the company’s future.

It couldn’t be learned whether the company’s struggles with the 4680 prompted Musk to hit the brakes on Tesla’s $25,000 vehicle. Musk hasn’t revealed whether the robotaxi will use the 4680, though analysts believe it will employ a different, conventional battery because the autonomous vehicle must withstand more punishment since it functions as a taxi.

In April, shortly after word of the $25,000 car’s fate leaked in a Reuters story, Baglino, who had been overseeing the 4680 battery, among other things, left the company. Tesla promoted Bonne Eggleston to senior director of the 4680 project, reporting directly to Musk, according to a Tesla org chart viewed by The Information.

Immediately after the Reuters report, Tesla’s shares began to drop, sliding 14% by the time Tesla next held a call with Wall Street analysts, on April 23. In that call, Musk didn’t promise to resurrect the cheap car, but said he would accelerate the launch of other unnamed “affordable” models. He promised to share more information about the vehicles in August, when Tesla was also planning to reveal the robotaxi.

Lars Moravy, Tesla’s vice president for vehicle engineering, whom Musk also assigned to help supervise the 4680 and dry cathode processing, meanwhile told analysts in the company’s first-quarter earnings call that he expected to beat suppliers’ 4680 cost estimates by the end of the year.

Now Tesla’s 4680 team is under the gun. In May, senior executives gave them the end-of-year deadline to make dry coating work at high speed, and to bring the battery’s cost equal to competing estimates submitted by at least one supplier, according to sources familiar with the matter.

The 4680 team aims to validate that its engineering approach to high-volume production works on a manufacturing line by September, according to a person with knowledge of the situation.

First though, he wants them to scale up a fix the 4680 team found to a flaw in the 4680 that often causes the core of the batteries—the electrode sheets, wound into the shape of a cylinder, then inserted into the 4680 metal can—to collapse in on themselves. Such a mishap can ruin the battery.

Musk ordered the team to produce a one-off 4680 battery pack containing both dry cathodes and a resolution for core collapse by this month to go into a Cybertruck that Musk wants to publicly show, according to a person familiar with the matter. Earlier this month, the team successfully produced this one-off battery, the person said. Now it must figure out how to make the 4680, with the fixes, fast and at large volume.

Some members of that team are skeptical they’ll be able to hit the year end deadline. Musk, for his part, seems to have started downplaying the importance of big battery breakthroughs.

At Tesla’s annual shareholders meeting last month, one investor asked Musk whether 4680s would make the company’s cars more affordable. Musk responded that battery prices had fallen so much over the past year that getting to cheap EVs was now much less about cheap batteries than it once was.

“Achieving affordability is actually at this point more about the rest of the car than it is actually about the battery pack,” he said.