The Crux of the Manufacturing Problems
Intro
Much has been made of transcripts from a “disgruntled ex-employee” of Enovix. We think his views are a valid data point in understanding the issues Enovix faces, whereas some longs dismiss him alleging he’s biased and uninformed. Let’s dig deeper into the truth.
Why Does It Matter?
Enovix isn’t an active supplier in the consumer electronics battery landscape, and little is known or discussed about the company outside of the retail stock trading community. Getting an informed perspective on Enovix’s technology from seasoned battery experts is somewhat hard to come by. This, in and of itself, is a red flag – if Enovix’s technology was so revolutionary, competitors would be familiar with it and there would be more discussion about Enovix in industry circles. Instead, the hype around Enovix’s batteries comes mainly from management, its stock promoters on FinTwit, and sellside analysts angling for fees from future capital raises. Naturally, these parties are biased to be bullish because they benefit when retail investors buy into the story and drive the stock price up. It’s a game that works until it doesn’t, and we hope there are longs that care about whether the company has a legitimate operational future, as opposed to treating ENVX as a trading sardine in a pump-and-dump scheme.
Given the lack of broadly publicly available views on Enovix’s technology and manufacturing process, institutional investors sometimes turn to “expert networks” to try to better understand a small company’s technology and business proposition. Through speaking with these “experts”, they can get independent insights from unbiased, or at least less biased, individuals, versus the more biased views from management, sellside analysts and stock promoters.
How Do Expert Networks Work?
There’s numerous expert networks, and examples include Gerson Lehrman, Guidepoint Global, Tegus, Stream, etc. These expert networks find industry experts and schedule calls between them and institutional investors, where the institutional investors can spend an hour or two asking the experts questions about a company or industry. The expert typically gets paid a fixed amount, like $500/hr or $700/hr. The institutional investor is charged a fixed amount, like $1000/hr. And the expert network takes a cut for finding the experts and arranging the calls. Everyone wins – the expert gets paid an hourly fee for sharing his or her opinions. The investor learns more about his subject company or industry from oftentimes unbiased industry specialists. And the expert network makes a fee.
With Enovix, battery experts are only moderately helpful – many just don’t know much about Enovix’s technology, or only have a perfunctory knowledge of what the company is trying to do. The most useful “experts” end up being former employees, because by actually working at the company, they get a direct view into what the company’s goals and business plans are. That’s intuitive – we all know much more about our own specific employer than any of our employer’s competitors, or the overall industry our employer operates in.
Are these employees sometimes biased? Certainly. Maybe they hated their time at the company. Maybe they were in a role that didn’t expose them to important and relevant knowledge about the company’s overall prospects. That said, usually the experts just answer the questions asked of them truthfully, given that regardless of what they say, they’re getting paid their hourly fee for speaking with the institutional investor.
Some of these expert networks transcribe these conversations and make them available to all paying members of the expert networks.
With Enovix, there haven’t been too many former employees that the expert networks have successfully convinced to speak with the network and allow the conversations to be transcribed. By our count, across three platforms that we looked at, we found 8 transcripts from former employees. Those 8 transcripts were actually interviews with only 3 people: the same former employees did multiple interviews. And 4 of the transcripts were with a fellow who the longs have outed, and who they keep referring to as a “disgruntled ex-employee”. We won’t mention his name here, but it’s become publicly discussed who he is. In this discussion, we’ll call him ”Gerald”. Of the two others interviewed, both left in 2018, so their views are a bit dated, though helpful as well. But Gerald’s discussions, in his 4 transcripts, have been particularly detailed and colorful, and my personal perspective is that he’s sharing his genuine opinions about Enovix. The longs have painted him as being biased against Enovix. While we can’t disprove that, he says many nice things about Enovix and its management in his interviews, and he just doesn’t come across as vindictive or angry. Rather, he comes off as relatively informed about the problems Enovix has faced. So I do think it’s informative to share his perspectives and why it matters if you’re long Enovix.
Gerald’s Background
Gerald worked with Enovix’s founder and former CEO, Harrold Rust, at IBM before Rust founded Enovix in 2007. After IBM, he worke dat a variety of companies, including Apple, taking on increasingly senior roles in manufacturing facility and process buildouts and enhancements. He left Apple in 2019, and then Rust reached out to him to convince him to come to Enovix in March 2020 and work on Enovix’s manufacturing buildout. He lasted for 11 months, and left in January 2021, to work for Google.
So, yes, he wasn’t at Enovix forever, and he’s been gone for two years, and hasn’t been there during the current management team’s tenure. But his insights are relevant, nevertheless, in our opinion. We’ll go through his insights, gleaned across the 4 transcripts. The following discussion is based on Gerald’s views, as represented though his expert network calls.
What Enovix was Doing in 2020
When Gerald joined Enovix, the company was building battery cells manually. “Yields” were very low (as in below 10%). Yield refers to the percentage of finished batteries that are fully usable and functional. So at that time, Enovix was building battery cells manually and a lot of them didn’t work when completed. There were two sizes of cells at the time, and some of the completed working batteries had capacities (ie. energy per volume, as measured by “watt-hours per liter”) that were 20% higher than the industry standard. So there were working batteries being completed, and those working batteries had higher capacity than the industry standard, but that higher capacity was only 20% better. And, again, a lot of the finished batteries weren’t functional (ie. the yields were low). And all the batteries were being manufactured manually. So if Apple needs 1 million batteries per day in order to be able to put Enovix batteries in the iPhone, Enovix wasn’t exactly set up to provide that.
Enovix’s technology, as discussed extensively, involves “stacking” silicon anode, as opposed to the conventional “jelly roll” winding of the graphite anode. This new “architecture” is novel. The problem is that existing battery manufacturing machines don’t exist to build batteries by stacking anodes.
These silicon anode strips are stacked on top of one another, in Enovix’s architecture. Each strip is very very thin, and hundreds are stacked on top of one another to make an Enovix battery. And they need to be stacked on top of one another at tight tolerances – in other words, there needs to be very little space in between strips, and then each strip needs to be stacked exactly on top of one another, not a little bit to the side. So to make a functional battery that achieves the 20% higher capacity that Enovix boasts, hundreds of extremely thin strips need to be stacked exactly on top of one another, very tightly. It all has to be very very straight.
Unfortunately, conventional equipment can’t be used to make Enovix’s cells, because the company’s stacking methodology is unique only to Enovix. So new equipment has to be designed. Knowing that this equipment design can’t necessarily be repurposed across the entire battery manufacturing industry, and suspecting that Enovix will never reach scale, equipment manufacturers charge Enovix very high prices to create this custom equipment.
To make money, Enovix in its own SPAC presentation projected to make 45 million cells a year. To make 45 million cells a year, you need a lot of equipment. But because of the highly custom nature of Enovix’s manufacturing equipment, the equipment is very expensive. To justify that cost, you need significant customer orders. But to provide committed customer orders, customers need certainty of lots of supply. Essentially, Enovix faces a significant chicken-and-egg problem. In fact, all startup consumer electronic batteries face this problem, which is why giant Asian conglomerates like ATL and Panasonic make the vast majority of consumer electronic batteries.
What Exactly is Enovix’s Value Proposition?
Enovix’s best performing batteries were merely 20% higher capacity than the industry standard, back in 2020. (Since then, as peer technologies have likely advanced faster than Enovix, it’s very possible that this capacity premium is lower, if it even exists at all anymore). By the way, ask yourself whether Apple, Samsung, Dell and other large customers would pay much for 20% longer battery capacity or battery life? We don’t think they would.
Now, the bigger problem is that Enovix is nowhere close to being able to manufacture the massive number of batteries that any of its customers need. During sales discussions when Gerald was at Enovix, customers would tell Enovix that if it built the batteries, they’d take a look at them and test them. But no one was interested in giving Enovix committed purchase orders. Without purchase orders, Enovix can’t justify the large investment to purchase all of the very expensive equipment required to build the large number of batteries that its customers need. And its customers aren’t desperate for Enovix’s batteries – a 20% higher capacity battery isn’t essential, or valued especially highly, by its customers. In fact, would customers pay 20% more for a 20% higher capacity battery. The general feedback was that most customers wouldn’t pay more. Companies like Apple, Samsung, Dell and Garmin won’t spend a cent more than they need for much of their supplies, which includes their batteries.
Then, keep in mind the timeframe around how a customer qualifies its vendors. These large customers take anywhere from 18 months to over 2 years to qualify and test a module. With batteries in particular, given the safety component, the testing and qualification process is particularly intense. For most customers, taking a flyer on Enovix simply makes no sense.
Yield, Speed and Cost
Another problem that Enovix faced is that in order to build enough batteries to be profitable, Enovix had to make 45 million cells a year, or 120k cells a day, assuming 365 days a year. But given the complexity of stacking large amounts of extremely thin strips almost exactly on top of one another, the machines had to operate at very slow speeds in order to achieve high enough yields. Unfortunately, these very slow speeds would make it very difficult to build the number of cells that Enovix needed to make each day… unless Enovix bought a lot of machines. Remember again that the cost of these machines, because they are so customized to Enovix’s architecture, is very expensive. So Enovix needed to buy a lot of very expensive machines. Again, the cost of production underlying Enovix’s business model is far too high for Envoix to ever be a legitimate, competitive player in the consumer electronics battery landscape. Yes, Enovix, in 2020, made a battery that had 20% higher capacity, in a lab setting. But the path to commercialization made little sense. It couldn’t produce these batteries at a high yield. The machines to make the batteries were going to be very expensive. Its customers weren’t going to pay much of a premium for these mildly higher capacity batteries. To be profitable, Enovix needed to produce a massive amount of batteries.
Finally, Enovix needs to make multiple sizes of batteries. Well that requires for all the custom equipment to be tuned to the different sizes. And it can take weeks of downtime to tune these manufacturing equipment to build batteries of different size.
Who Designs and Builds the Machines
When Gerald was at Enovix in 2020, Enovix turned to DW Fritz to design the manufacturing equipment and process. Unfortunately, DW Fritz overpromised (to win the contract), then underdelivered.
Enovix needed a variety of machines. There was a machine line for laser cutting for the anodes, one line doing laser cutting for cathodes, one line doing the laser cutting for the separator, a stacker machine, a machine for the overfill, a machine to install the constraints, a tearing machine, and others.
To reiterate, no other battery company employs Enovix’s sort of stacking architecture, so these machines haven’t been designed before.
Turnover
Gerald discusses that many engineers who he believed were talented left during his tenure, and others he’s been in contact with have left since. In fact, it’s instructive to search LinkedIn and go through who on Linkedin works at Enovix. We came away unimpressed by the overall number of quality engineers currently working at Enovix.
Management
Gerald had a high opinion of Harrold and viewed him as intelligent, competent and honest. But he felt that upon the SPAC acquisition completing, Harrold came under a lot of pressure from Rodgers to perform. Unfortunately, the specifics of Enovix’s technology makes it impossible, in our opinion, to actually commercialize Enovix’s technology, regardless of who the CEO and COO are.
Conclusion
While Gerald was at Enovix in 2020, the company could occasionally produce some good cells, and the good cells had 20% higher capacity than the competition. But they were very expensive to manufacture, and the yield was low. The result was that Enovix had no real believable pathway to becoming a competitive player in the consumer electronics battery landscape, so Gerald left, and so did many of the other engineers.
This discussion around the difficulties around taking Enovix from lab to competitive commercialization explains some of what we’ve seen as investors. It explains why during the SPAC acquisition, the company projected 45m cells/yr and $220m of revenue at FAB-1, but now is projecting $1m of revenue. Instead of 45m cells/yr, its 2Q23 forecast is now 18,000 units (which they’ll probably beat!). 18k units at $10/cell is $180k of quarterly revenue. Enovix’s valuation is north of $2b. It’s all a little silly, to be honest.