Aehr Test Systems (AEHR) New Customer & CFO

Investment Analysis Clubs Saul’s Investing Discussions
21

Actually, if we want to look growth rates then we need to consider that Toyota makes just two BEVs today - the bZ4X, which is also sold as the Subaru Solterra, and the new Lexus RZ, which was just introduced in March this year. The Lexus RX is the first Toyota vehicle to use SiC in its inverters. What’s ironic there is that Toyota has been looking at using SiC in its vehicles since at least 2014, but apparently did not bring the tech to production vehicles until just now.

It does seem that all OEMs are moving towards using SiC in their inverters, replacing the older (cheaper and less efficient) IGBTs. For instance, here’s an article on Hyundai delaying their Ioniq5 BEV because of IGBT supply problems.

Hyundai is looking for alternative options to secure automotive chips from other semiconductor companies, including Switzerland-based STMicroelectronics, to prevent production disruptions to IONIQ 5.

I wouldn’t be surprised to find out that the switch to STMicro isn’t for IGBTs, but for STMicro’s SiC chips, which as I pointed out earlier, is what Tesla has been using since 2018. And again, STMicro is a big customer of AEHR.

AEHR is a really interesing picks and shovels play for EV adoption, but there is always the potential for newer inverter tech to come along. I don’t believe there’s anything that’s going to happen soon nor fast, though.

18 Likes
22

Smorg, where do you see evidence that STMicro is a big customer of AEHR? That would change things for me, but everything I’ve read indicates that 80%+ is from Onsemi and that will only increase for at least the upcoming quarter. Is it an upcoming deal or just wishful thinking from the sites you’ve read?

Scanning the last few 10-Ks it seems AEHR has a pattern of keeping large customers for a year or two, losing them, and then replacing them with another large customer. They’ve listed Texas Instruments (in 2019), Intel, ASE, and yes STMicro as past customers.

Directly from their 2022 10-K (5/31 year end)

During fiscal 2022, ON Semiconductor accounted for approximately 82% of the Company’s net sales. During fiscal 2021, Advanced Semiconductor Engineering, Inc., ON Semiconductor, Intel and Inphi accounted for approximately 24%, 23%, 20% and 10%, respectively, of the Company’s net sales. During fiscal 2020, Intel, ON Semiconductor and STMicroelectronics, accounted for approximately 43%, 16% and 15%, respectively, of the Company’s net sales. No other customers accounted for more than 10% of our net sales for any of these periods.

Their disclosure on this in the next paragraph:
We expect that sales of our products to a limited number of customers will continue to account for a high percentage of our net sales for the foreseeable future. In addition, sales to particular customers may fluctuate significantly from quarter to quarter.

No doubt that’s lawyer speak/cookie cutter SEC disclosure, and like I said I think the stock will do well, with possibly a 60% QoQ revenue increase for 5/31, but my big hangup is the reliance on Onsemi, and if that’s truly not the case in the foreseeable future, then I would look to make room in my portfolio for AEHR

11 Likes
23

Good catch, @Aphalite. Turns out I was looking at references to older docs, and made the mistake of just reading the search summary instead of opening the actual file. That said, the percentages vary all over the place, as AEHR points out sales will fluctuate from quarter to quarter.

Given that AEHR’s business is growing, I’m not surprised that their top customers are changing positions relative to each other, and don’t necessarily see that as losing customers as it could be other customers are simply buying more. The reporting in the 10-K is the standard ‘list any company which is more than 10% of the business’ disclosure, and sure, those are risks to consider. That said, I would still consider STMicro at 15% of AEHR to be a “big customer,” as I said - just not AEHR’s biggest customer.

So, I guess the question is whether STMicro is going to scale AEHR usage up or down? But, even if down, it still seems to me that as more and more EVs are sold, more inverters are needed (and even hybrids need inverters), and by far the best technology choice for inverters are SiC chips, so AEHR’s SiC burn-in/testing TAM is growing. It would be great to have some insight into what STMicro is thinking here, but I don’t know how to gain that insight.

7 Likes
24

Given that AEHR’s business is growing, I’m not surprised that their top customers are changing positions relative to each other, and don’t necessarily see that as losing customers as it could be other customers are simply buying more.

Would agree with this if the percentages weren’t so lopsided in 2022 and 2023. Something like the 2021 spread or even the 2020 spread would be great. But I understand everyone has their own risk tolerances when it comes to customer concentration.

That said, I would still consider STMicro at 15% of AEHR to be a “big customer,” as I said - just not AEHR’s biggest customer.

That doesn’t seem likely to me. The last time STMicro was disclosed as a large customer over 10% of net sales was in 2020, when sales was 22.3m. At that point 15% of sales would be 3.4m. In 2021 STMicro was not a customer over 10% when sales was 16.6m, which means if they were a customer they contributed less than 1.7m. It is possible they were still less than 10% last year, when the cut off was $5m to hit the threshold, but it seems to me a customer like STMicro (COGS of $8 billion+ in 2022), if they chose to continue to use AEHR’s solutions and with SiC apparently growing (I say this because it doesn’t seem clear to me this is true either, with Tesla’s recent announcement) in market share, would easily clear the $5m hurdle. Of course, things could always change in the future, just thought you maybe had some additional insight into your claim as a relationship with STMicro would directly relate to the Toyota announcement.

8 Likes
25

Actually, it looks like Toyota’s inverter supplier, Denso, might even be making its own SiC chips:

DENSO has developed a new in-vehicle SiC transistor, and this marks the first time DENSO has used SiC for in-vehicle diodes and transistors. The newly developed SiC transistor offers both high reliability and high performance in in-vehicle environments, which can challenge semiconductors, thanks to DENSO’s unique structure and processing technique, which apply trench gate MOSFET.

I do not believe Denso is an AEHR customer, but obviously don’t know for sure.

6 Likes
26

Separately, here’s an article forecasting growth in SiC at a 36.4% CAGR for the next 5 years:

The global silicon carbide market was valued at USD 1.8 billion in 2022 and is projected to reach USD 11.1 billion by 2028; it is expected to register a CAGR of 36.4% during the forecast period.

Automotive segment is expected to experience the highest CAGR of 38.2% during the forecast period. The rising adoption of EVs/HEVs and electrified powertrains will boost the demand for silicon carbide devices.

And, as with all things EV, I suspect volumes will be greater than forecast. This doesn’t directly translate into more business for AEHR, just that their potential market is growing well.

10 Likes
27

Are you certain that these customers are “lost?”

I just can’t imagine that a customer would spend a few million dollars to buy AEHR’s testing machine and a few million more to purchase the machine that aligns the wafers on the test tray so that the process can be automated. Go through whatever factory mods that were necessary to install the machines. Get their industrial engineers (or business process experts, or whatever they call them) to redesign the business process, train up their people on the new process and work out the kinks over the first few months only to scrap the whole thing and decide that it was a blunder after a year or two.

That just doesn’t make much sense. It seems pretty evident to me that installing AERH’s machines is a long term commitment.

I think it’s more feasible that they bought a lot of the expendable components (razor blades) with the original purchase contract and simply didn’t need to place any new orders for an extended period of time. I read the materials AEHR provides on their website, which isn’t very much, but I got the distinct impression that they don’t build inventory. They build every machine to order and there’s a limited amount of customization available to satisfy the unique requirements of each customer. In other words, the razor blades are not interchangeable from one customer to the next, they are built to match the characteristics of the customized machines that were installed in that customer’s factory.

If that’s the case, I would imagine that the customer would probably buy a lot of razor blades with their initial order, and it’s likely that AEHR gives them a special price break on razor blades purchased at the same time the machine is purchased. I don’t think the absence of orders for some period of time is indicative of a “lost” customer.

Yes, I am speculating about this, but if you consider the initial investment required in order to become an AEHR customer in the first place it just doesn’t make sense to then throw it away. And I think it even more unlikely that customer after customer would do this. If that were the case, if there was some discovery after using the AEHR machines and all, and customer after customer made the same decision to scrap it after a couple of years, how on earth could AEHR stay in business?

They claim on their website that they have over 2,500 systems installed all over the world. The company was founded in 1977. If you google AEHR competitors, four companies are listed. They are: Advantest, Teradyne, Form Factor and Tokyo Electron. I’ve not researched these companies in order to determine if they are truly competitive when it comes to cost and scale. I doubt it in that AEHR is pretty blatant in declaring that they have no competitors (listen to the William Blair webcast).

9 Likes
28

I definitely can’t say for sure if the customers are lost or not, but lack of large orders for a long period of time sure does seem to me like the same thing. Those listed customers all run through billions in costs each year, although a few million might seem like a big deal in AEHRs case, it’s miniscule in scope to the scopes of Intel and STMicro. If they loved the product, why are they not putting in multi million dollar orders like Onsemi? Surely this increase in SiC demand would mean higher need for testing for them too?

All that said, I’m waiting for the 2023 10-K (should be in August) to see if the disclosures match the claims from the conferences. Maybe the product is just slow to catch on and that’s why not so many customers have placed orders yet and they’re just in the discussion phase. One nice thing AEHR has going for it as an investment IS the small size. As Saul has already pointed out, you can grow much quicker from small orders when you only run $15m in revenue quarterly to begin with.

As far as all of those competitors that you listed, I did a quick Google search and they all have billions of dollars in revenue, with AEHR having been in business 40+ years, it sure does seem like if they really had such a great product, their business should be much bigger. The CEO certainly seems enthusiastic, but you’d definitely want that type of enthusiastic messaging at a public conference, until they convince more customers to sign on (or start ordering again, if they are just dormant as you say), I think the revenue concentration is a huge risk

17 Likes
29

I used to work at Boeing. They had more revenue than a lot of countries had GDP. I can assure you that though a few million might seem “miniscule,” any factory manager who squandered a few million would be looking for a new job. I don’t know what caused the apparent large customers failure to place new orders, may be that the demand for SiC was being met by On at a price point that made it unprofitable for other vendors to compete at this time. Could be that they have just postured themselves to enter the market quickly when demand spikes. Just speculation. I don’t know???

Yes, AEHR has been in business for 40 years give or take, but they have not been offering these SiC test machines for that long. I think the AEHR rep at the Blair conference said the patents are 3 years old, not sure about that, but somewhere I got the understanding that the patents are relatively new. If 3 years is the right number it means they’ve got another 17 years before the technology becomes public domain. It also might explain why AEHR isn’t knocking out big contracts on a regular basis. But they do claim to have an installed base of 2,500 systems worldwide. In addition I am pretty sure they guy at the Blair conference said that AEHR will have a bunch of new patents for GaN wafers soon.

But I admit, I really only have speculative answers to address the issues you raised. Facts are facts. They have only told us about the orders they have publicly announced. Even if a lot of customers don’t want their orders to be public, the announced orders account for a large percentage of their revenue.

So yeah, you can wait for the answers and pay double or more for the stock. Or take a victory lap when the stock price tumbles.

9 Likes
30

I’m not rooting for the stock price to tumble, in fact, I think the most likely scenario here is that the stock will keep climbing, ON seems to very much love AEHR’s solutions, and the orders appear to be quite consistent. It’s just that for my own personal risk tolerance, a single large customer automatically disqualifies a potential investment only because I’ve seen the worst case scenario happen before. Now AEHR is a little more diversified than that, as ON is only 84% of revenue and the company has survived a changing of large customers before, so I’m not drawing parallels, only explaining why it’s scary.

But I’m not trying to project my risk tolerance onto you or anyone else that invests in this stock. +43% in the past month and +127% YTD is something to celebrate - I certainly don’t have results like that. And in the end, no one will remember in hindsight if there was massive risk taken to get there, they’ll only see your results and applaud

11 Likes
31

CEO seems like very sincere, likable guy who has spent his life in the semiconductor industry, saw something special in AEHR a decade ago (he tried to acquire it for HP where he was working but his company was spun out, renamed Verigy and itself acquired.) He then takes over AEHR and brings a strong vision for product development - which he clearly knows well. While at AEHR he builds products based on where he sees the market for his products heading. He knew he was early and apparently was dead on with his vision, execution. Took roughly a decade to pull it off but he did. Impressive story. And clearly he had Board’s support for a long term vision. This really does show how it’s done - patience, vision, working on things aligned with his expertise - no growth for first 8 or so years as foundation laid, then when world ready, he’s there to meet demand and stock price rockets from low single digits over 40 as of this writing.

Somewhat strangely here’s a video of him running to join board of his country club…

Again, seems like very nice person who favors getting things done without animosity. My one thought would be is he the right guy to lead charge going forward now that vision realized? That he’s thinking of “rolling up his sleeves” to help drive membership at his country club doesn’t scream hungry to conquer the business world to me. But obviously that’s pure speculation. And CEOs of course entitled to life beyond the job which may be positive in providing balance. And suspect he was once “Student #1” in comedy film “Campus Man” shot at his alma mater Arizona State. (Just trying to humanize the companies folks, since, you know, humans run these companies.)

Bottom line on CEO he had the right vision, the patience to realize it and deserves great credit for that. I love to see CEOs who have spent their lifetime in their respective industries - leaders like George Kurtz at CRWD, Jeff Green at TTD and Eric Yuan at ZM. You can’t fake decades of work. Clearly he knows and loves semiconductors.

9 Likes
32

I agree, and this is a legitmate risk. The narrative counterpoint is twofold:

  1. Management has confirmed two new customers with a chance to scale similarly to Onsemi IF test burns are successful. While the tests aren’t in the bank, the potential size of the commitments are. This new business would certainly lessen concentration risk IF (there’s that word again) Aehr executes.

  2. The rapid growth of the EV market could play right into AEHR’s hands, especially at such a small scale. Much like Jeff Green’s bet on programmatic ads seems to be paying off for TTD, Aehr’s bet on SIC looks like it could be a significant tailwind.

Every company is a mix of what could be versus what is. Aehr is no different. In this case, I am betting Aehr’s “could be” has a chance to show up on a much faster timeline with bigger numerical impact than say AI’s potential effects on NET or SNOW.

All the above is still obviously TBD. Regardless, this is a good discussion of Aehr’s possible risk/reward.

16 Likes
33

Thanks for that insightful and humanistic post - And welcome back Dan, you’ve been absent (or at least not posting) for quite a long while.

6 Likes
34

During the next conference call I wish some analyst would ask point blank,“We see that On has 84% of your business. What’s the story with your other 2,499 customers?”

8 Likes
35

This is a long thread already but thought to add two things:

  1. an analyst did ask point blank what the future holds given the backlog trends. The CEO answered that they don’t give quarterly guidance and then said this, in which he addresses customer concentration going forward:

“The reality is that, I see things picking up. There seems to be a lot of people accelerating their plans right now. We are having – we are putting out a lot of quotes. There’s a lot of people asking for lead times on top of each other and I believe we will go into next year with a lot more visibility than even last year.

Having said that, it will be distributed across a much larger number of customers, which I guess there’s some averaging going on in terms of that – it’s not all or nothing with one or two customers, but we also think that the large customers are going to continue to buy.

So I don’t know how many different ways I can describe our optimism, but Aehr is – I mean, people are really recognizing the value that we add and are seeking us out in addition to us knocking on all the right doors and I am super enthusiastic about heading into next year.”

  1. the lumpiness of revenue and implicitly why customer concentration can spike like it did despite what looks like a groundswell of customer demand. I believe it may be due to their revenue recognition policies. CEO again on the last call:

“Our policy, I think, is very conservative. If we have a new product, particularly to a new customer, but if we have a brand new product that has never been proven or installed and accepted by the customer, we simply don’t take revenue for it until that milestone, even though we know it’s working here, it’s been completely proven out, et cetera, but until the customer actually signs off on it, we won’t score revenue recognition.”

→ I think the customer concentration argument could therefore appear larger than it has been historically, and would certainly seem to be heading in the right direction going forward.

-wsm

31 Likes
36

To rephrase another way: “we see that two customers have 94% of your business, and three more customers have another 5% of your business.
What’s the story with your other 2495 customers who altogether comprise 1% of your business?”

Two customers accounted for approximately 82% and 12% of the Company’s net sales in the three months ended February 28, 2023. Sales to the Company’s five largest customers accounted for approximately 99% and 97% of its net sales in the three and nine months ended February 28, 2022, respectively.

4 Likes
37

There’s no way there are 2497 SiC customers out there for AEHR.

AEHR holders need to get over customer concentration risk or not be in this stock at all.

I get that they still need to penetrate more of the SiC wafer manufacturers however even if AEHR snags ALL of them there might be only 10 globally of any significance. It wasn’t long ago when it was basically just CREE, (Wolfspeed) and less than a handful of others.

There’s probably between 6-10 major full SiC manufacturers so I would expect AEHR to have to report their entire SiC customer base in their 10% customer concentration risk assessment for the foreseeable future. This might grow in the future as Wolfspeed and others supply other chip makers with the SiC raw material for them to use.

GaN is also relatively concentrated into a number of high volume players.

This is NOT the same as the Si chip manufacturer equipment market. So unless AEHR finds a way to compete in Si then I would focus on customer acquisition and penetration rather than concentration for AEHR.

Ant

41 Likes