Aehr is a different kind of company


I see concerns like the above about Aehr, but Aehr is an entirely different kind of company than the ones we usually invest in. With Aehr we are investing in the future of the company, not in the present. I truly didn’t even pay any attention to the above figures about quarterly revenue and earnings. What mattered to me was this kind of information (slightly shortened by me), see below :


“In this quarter just completed, we received the first purchase order from a new silicon carbide semiconductor company for our production FOX-XP solution to be used for volume production wafer level test and burn-in of silicon carbide devices for electric vehicles, trucks, and train traction inverter modules.

The train traction inverter application represents an exciting new market for our FOX production test solutions due to the extreme reliability and length of service requirements of this application leading to prolonged test times. This new customer, a multinational industrial conglomerate and manufacturer of semiconductors including power semiconductors, is forecasting to grow their silicon carbide business significantly to meet the market demand.

“With the addition of this latest customer, we have significantly expanded our customer base by adding a total of four new silicon carbide customers this year. Each of these customers is already ramping or plans to ramp our products into high-volume production using our multi wafer test and burn-in systems.

“We also have multiple potential customers inquiring about our systems with the new high voltage option to test and burn-in gallium nitride (GaN) semiconductors for power conversion applications. The gallium nitride market appears to be a potentially significant growth driver for our systems and WaferPak full wafer Contactors, particularly for automotive and photovoltaic applications where burn-in appears to be critical for meeting the initial quality and reliability needs of those markets.

“We also see a major market opportunity with silicon photonics integrated circuits for optical chip-to-chip communication.

This is in addition to the current photonics transceiver market used in data and telecommunications. Multiple companies such as Intel, nVidia, AMD, TSMC, and Global Foundries have made announcements…

During the fiscal fourth quarter, we received our first order from a current major silicon photonics customer for a volume production FOX-XP configured to enable cost-effective production test of wafers of next-generation photonic integrated circuits, which can be used in new optical I/O or heterogeneous integrated packages. This customer is one of the world’s largest semiconductor manufacturers and we expect to receive orders for additional production systems as they increase production of these devices."


In other words the number of areas their tech can be used is exploding, their TAM is exploding, their number of large customers is exploding. Who in the world cares whether their revenue came in at $22.0 million or $22.3 million? It’s totally irrelevant to the investing thesis.

For disclosure, Aehr was my 2nd largest position, at 15.4%. That’s before figuring in the 17% that they are up today (as I write).

Saul

78 Likes

I’ve read through most, if not all the commentary about Aehr on this board along with a great deal of what’s been published on SA and elsewhere. I don’t consider myself an expert on SiC applications, but I think I’m reasonably well informed.

From what I’ve read on this board it appears that customer concentration is considered by some folks as a major risk. And for a lot of companies in a different line of business I would agree, but it does not worry me for Aehr. The reasons it doesn’t worry me are for one, I believe Aehr’s largest customer is Onsemi. If you Goggle "which company is the largest producer of silicone carbide semiconductors you’ll find the Onsemi is among the top five. The other four are Alegro, Infineon, Rohm and STMicro. Of these, I’m quite certain that STMicro is another Aehr customer. I’m not certain about the other three, but if they aren’t current customers, I’m certain that they have at least been approached by Aehr.

If you Google “how many SiC chip makers are there?” you don’t get a simple answer, but after going through several returns, I’ve come up with eleven major manufacturers. In other words, if every SiC chip maker was an Aehr customer they would still have a relatively few customer. Even if eleven isn’t the right answer, maybe there’s 25 - I don’t know, but they will never have hundreds of customers, at least not for their high end $2.5M test systems.

Aehr has not revealed how many customers they have, but they have revealed that they’ve shipped 25,000 test systems (obviously, not all of these are the high end equipment). They have also said that two of the top four SiC makers are customers and a third is going through test and evaluation at present. Aehr does not book revenue until the customer accepts the equipment.

So if you need a high customer count to feel comfortable with an investment in a company, Aehr just isn’t a good fit for you.

41 Likes

I agree with Saul’s points and, @brittlerock while I also agree about customer concentration being less of a worry, I disagree slightly about the reason why. I believe we have very concrete reasons to not be worried about it going forward. That is because they actually have told us quite a bit about the number of SiC customers they have - by my calc they now have 5 - and they’ve also told us quite a bit about their large customers in years past, and going forward their customer concentration should fall rapidly. That’s the basis for my expected revenue calc which I did here. Below is the supporting bits for the customer numbers.

First, their disclosures about large customers

Q4 2023: Their largest customer (OnSemi) accounted for 79% of revenue - $51.4m
2022: One customers (Onsemi) accounted for 82% of revenue - $41.7m
2021: OnSemi (this was before they started ramping SiC) was 23% of revenue - $3.8m

From the 2021 10-K:

Advanced Semiconductor Engineering, Inc., ON Semiconductor Korea, Ltd., or ON Semiconductor, Intel Corporation, or Intel, and Inphi International Pte. Ltd, or Inphi, accounted for approximately 24%, 23%, 20% and 10%, respectively, of the Company’s net sales. During fiscal 2020, Intel, ON Semiconductor and STMicroelectronics, Inc., or STMicroelectronics, accounted for approximately 43%, 16% and 15%, respectively, of the Company’s net sales. During fiscal 2019, Intel, Texas Instruments Incorporated, or Texas Instruments, Cypress Semiconductor Corporation, or Cypress Semiconductor, and STMicroelectronics accounted for approximately 36%, 14%, 12% and 10%, respectively, of the Company’s net sales

→ So we know they have the who’s who of the market as customers, but they didn’t contribute that much to revenue. The big change came when OnSemi started rapidly ramping up SiC production and became their biggest customer by far in 2022.

Moving on to SiC customers specifically

I’m quoting the CEO in the different quarterly calls below. It paints a fairly clear and strong picture of growing demand and customer wins for SiC. I’m starting the quotes from Q4 2021.

Q4 2021

This year we successfully took the initial order core and installed our first production capacity for silicon carbide devices

This new silicon carbide application with a Fortune 500 market leader in silicon carbide and power modules […]

Since our initial installation in January we’ve received multiple follow-on orders for additional WaferPaks from this customer

→ They got their first SiC customer installed in January 2021 and this customer then started ramping. This gels well with Onsemi’s announced plans and execution.

Q1 2022:

Our strong bookings include several sizable orders received over the past few months from our lead silicon carbide Test and Burn-in customer for our FOX-XP wafer level test and burn-in systems and full wafer, WaferPak Contactors to support testing silicon carbide devices for electric vehicles.

This Fortune 500 customer is a major automotive semiconductor supplier, and we continue to work closely with them to achieve their test and burn-in requirements and capacity needs.

We’re currently in detailed discussions with multiple major silicon carbide suppliers regarding their wafer level test and burn-in needs. […] We believe we will add several new silicon carbide customers over the next 18 months that will ramp into production on our solutions.

→ Onsemi’s ramping up; they’re starting to talk to many other existing SiC players and they expect to land a number of them in the next 18 months.

Q2 2022:

During this last quarter, we received a follow-on $19.4 million order for FOX-XP wafer level test and burn-in systems from our lead silicon carbide customer.

We’re currently engaged in discussions and/or evaluations with several other silicon carbide suppliers regarding their wafer level test and burn-in needs. This includes all of the current large silicon carbide suppliers, and a number of companies both large and small that intend to enter the silicon carbide market.

Later in the Q&A

So just for clarity, the industry, market forecasters that talk out there, generally refer to the largest suppliers in this space traditionally have been ST had the number one position, and then Infineon and Cree, now Wolfspeed are others. And then ON Semiconductor that kind of came out of nowhere, and people know that On Semiconductor is one of our top 10% customers, are generally thought of by most people as the current larger players in the space. In addition to that, you have Mitsubishi and ROHM that historically were kind of filled out the top six.

We have some level, if not a deep level of engagement with all of the above

So I mentioned six companies. There’s also think, four, maybe five more […] Part of the challenge is how do you bet on – which horses do you bet on. Our current strategy is we’re betting on all of them. We have the capacity and sales

→ So they are going after and in discussions with all of the top suppliers - ST, Infineon, Cree, Wolfspeed, Mitsubishi, ROHM, and they have OnSemi already and it is their biggest customer.

Q3 2022

During the quarter, four additional companies provided detailed wafer layouts to Aehr for their silicon carbide wafers as part of their evaluation of using Aehr’s FOX-XP systems.

Last quarter, we said that we believe we’ll add several new customers that will be ramping into production by the end of our fiscal '23, which ends May 31, 2023.

→ Still only 1 customer, but the discussions of the previous Q have now led to evaluations.

Q4 2022:

Our lead customer for silicon carbide wafer level burn-in made significant investments in their silicon carbide production throughout this past fiscal year.

We’re excited to see them continue their ramp, and we continue to expect significant additional system and wafer pack purchases from them over the next several years and to the end of the decade, as they strive to be a market leader – market leading supplier of silicon carbide devices.

Q&A:

There are others. In fact there’s a – reviewing the funnel with Vernon and the team. There is a significant number of players that are talking about getting into silicon carbide. Many of them have been in now announced that include large and small and candidly, many actually see large players that have not announced their plans that have approached us and have been talking to us about their very real plans for entering the silicon carbide market.

So it’s welcome to have all these new players coming in, because all of the players today don’t have a chance of meeting all the demand as they basically go from 1x to 25x, the output by the end of the decade.

→ They still have only 1 SiC customer which is ramping and planning to continue to ramp but they are progressing discussion with all the others and there are big new players planning SiC production - which they are talking to - but these customers have not yet announced anything.

Q1 2023:

We’re currently engaged or in discussions with almost all the existing and future silicon carbide suppliers.

Our lead customer for silicon carbide wafer level burn-in continues to ramp up use of our FOX-XP multi-wafer systems and WaferPaks, placing yet another significant order with us during the quarter.

→ Still have only 1 SiC customer but now are talking to all of the existing suppliers as well as all the future suppliers.

Q2 2023:

We are excited to have added two new customers for silicon carbide test and burn-in during the quarter. The first is a major silicon carbide semiconductor supplier that purchased our FOX-NP dual wafer test and burn-in system used for engineering and device qualification during the quarter, and after the quarter closed, has since placed their first orders for two of our FOX-XP multi-wafer systems for volume production test and burn-in of their silicon carbide wafers, including the order we just announced today. This company is one of the world’s largest suppliers of silicon carbide devices and serve several significant markets, including the electric vehicle industry, as well as other industrial applications.

We now have two of the top four silicon carbide market participants as customers.

→ They added two and now have 3 SiC customers, including two of the top 4 - Onsemi and another, new one.

Q3 2023:

During the quarter, our second major silicon carbide semiconductor customer moved from an initial FOX-NP dual wafer test and burn-in system of ours, used for engineering and device qualification to purchasing their first FOX-XP systems to be used for production test and burn-in other silicon carbide wafers.

→ They still have 3. But their second big customer - one of of the current top four SiC producers in the world customers has moved to production. Remember that when OnSemi started to ramp end 2021, their revenue to AEHR went from $3.8m in 2021 to $41.7m in the next year, 2022.

Q4 2023:

With the addition of this latest new customer, we’ve significantly expanded our customer base by adding a total of four new silicon carbide customers this year.

→ In Q2 they said they added two. Now they say they added four in the year, which means they added another two in this Q, and now have five SiC customers. - Onsemi and four others, of which one is in the top 4 in the world and has started to ramp.

And then they also told us:

We continue to work closely with one of the largest silicon carbide players in the world on a large wafer level benchmark and qualification for automotive and other markets.

While candidly this has taken much longer than we thought it would, we’re excited that this qualification continues to make good progress and we remain confident that it will result in them moving to our FOX wafer level systems and WaferPaks for their volume production. They have told us that their plan is to move all new production capacity to wafer level burn-in and away from packaged part burn-in.

→ So in addition to the existing 5, there is another big one waiting in the wings which is not yet a customer but is soon expected to be.

Putting all of this together

They have just about everyone as a customer for bits and bobs - they’ve been in business for 40 years so this is to be expected. But revenue has only recently started to explode due to SiC, and as of now only based on the rapid scaling of SiC production by one large customer - likely Onsemi. That of course led to revenue concentration from this one customer.

But during this last year they have added four new SiC customers for a total of five. And there is another big one waiting in the wings which is expected to become a customer soon. Standing back a bit, this seems like a water-shed year. The big revenue ramp of the last year and a bit was on the back of just one customer in SiC. And now they have five (and soon six).

And that is before even looking at photonics, which they expect will be bigger than SiC.

87 Likes

I‘m deeply impressed with the level of detail and the amount of effort you put into researching the companies. I‘m learning a lot from you all and I just wanted to say THANK YOU for your contribution.

I really appreciate what you all are doing here!

Cheers
Nick

27 Likes

First, I wanted to credit “bits and bobs”. I’ve never heard that phrase before, but it’s awesome and made me laugh.

I’d like to point out this exchange on the call regarding customer concentration:

CFO Kenneth Spink

We’re filing our 10-K at the end of August. I think August 28 is our proposed filing date right now.

Analyst Christian Schwab

Could you be willing to share if there was more than one 10% customer and exactly if there was only one – exactly how much revenue came from just the one customer?

Kenneth Spink

Yes, absolutely. We did actually have 2 customers over 10%, one at 79% of revenues and the other at 10% of revenues.

Christian Schwab

Excellent. And Ken, when you guys are talking about your greater than $100 million expectation for revenue next fiscal year, how many 10% customers are you assuming in that guidance?

CEO Gayn Erickson

Let me try and field that one. Maybe almost better I don’t have it in front of me because I don’t want to get too specific on it. But I think we expect several, maybe [ 3 or 4 ]. So there will be – we’re obviously very heavily focused on our big silicon carbide customer this last year. And we expect while they will continue to be a great customer and certainly a 10% customer for years to come that there will be less of a focus in terms of concentration with them.


Doing some quick-and-dirty math, that implies going from one customer at $51M and another at $6.5M this year to potentially 3-4 customers at $10M+ next year. As others have pointed out, there are only so many SIC customers Aehr can chase anyway. I’m personally much less concerned with customer concentration now than I was a couple days ago.

This. With the lack of IPO’s over the last 18 months, it has been darn near impossible to find small companies with the chance for true hypergrowth. While AEHR’s math is lumpier than most, it does appear to have that chance (and is already doing it profitably I might add).

Size your bets accordingly.

38 Likes

Not to beat a dead horse - but you’ve pretty much proved my point, “by my calc they now have 5.” Exactly, you had to do a whole lot sleuthing to come up with the number of customers. In other words, Gayn has straight out told us how many customers they have for the high end FOX system. I admit it, I haven’t done the same detective work.

If you think about it, we don’t even know how they define “customer.” When does a prospect become a customer? At the time they make a verbal commitment of some sort? When they sign a purchase agreement? When they place an order? When Aehr records revenue? We just don’t know.

And I don’t care. I don’t worry about exactly when a customer is recognized and I don’t care about the current number of customers irrespective of what the number might be.

To be totally honest, there’s very little that I find to worry about with respect to my investment in Aehr - maybe the percentage of my portfolio. I should probably sell something and put the money in Aehr.

19 Likes

Great post, Brittlerock, I love it !!!

Saul

10 Likes

When the check does not bounce.

Denny Schlesinger

21 Likes

They’ve sold/installed 2,500 systems (all product lines) since the dawn of time. In 1997 - by the time of their IPO - they had "sold more than 2,000 systems to semiconductor manufacturers, semiconductor contract assemblers and burn-in and test service companies worldwide". At the end of fiscal 2005, it was upped to “over 2,500 systems installed worldwide”, and has stayed at that “over 2,500” level since then.

Aehr had entered into an agreement with DARPA to develop a wafer-level burn-in and test system, and the full wafer contact burn-in and parallel test system was introduced in 2001 (FY02), along with the WaferPak cartridge system. (FOX-14, capable of testing up to 14 wafers at a time.)

Here’s the progression of revenue by product line looking “a few” quarters back.

There are a few gaps, and the data is somewhat approximate early on, but the cumulative revenue from wafer-level systems and services, including contactors, ends up at about $300 million.

Data for systems (all product lines) vs contactors only goes back to FY 2017. Here’s a complimentary chart with “a few” quarters of context:

A best attempt at combining the two gives the following, which necessarily includes services. (FY23 is approximate, since there’s no 10-K yet.)

Wafer-level
Systems & Services
Contactors Packaged Part
Systems & Services
Contactors % of
Total Wafer-level Revenue
2017 7.6 2.0 9.3 21%
2018 6.6 6.5 16.5 50%
2019 8.5 6.2 6.4 42%
2020 9.0 10.8 2.5 55%
2021 9.2 5.8 1.6 39%
2022 26.3 22.6 1.9 46%
2023 41.9 21.9 1.2 34%
Total 109.0 75.8 39.4 41%

(Cumulative revenue from wafer-level systems and services prior to FY 2017, including contactors, is somewhere around $110-$130 million.)

In other words, cumulative revenue for wafer-level systems minus “consumables” puts a cap on the number of installed systems (of interest). So it’s a lot less than 2,500, and I feel @wsm007’s excellent post is a good way to approach and think about it.


Customer concentration, highlighting Onsemi, currently looks like this:

The mix of five largest customers isn’t constant, so here’s a more detailed breakdown covering the last decade. (Note that a blank cell only means less than 10% of revenue from a customer. The column to the right provides the upper limit, i.e the lesser of 10% and what’s unaccounted for.)

ON Intel AMD Inphi STMicro TI Cypress Astronics Apple Micronas Spansion Max revenue from any unlisted top 5
2014 7.9 2.4 5.9 1.6
2015 4.5 1.1 1.0
2016 4.6 6.8 1.5
2017 3.2 3.6 8.5 1.9 0.4
2018 7.7 10.0 3.8 3.0
2019 7.6 2.1 2.9 2.5 1.7
2020 3.6 9.6 3.3 2.2
2021 3.8 3.3 4.0 1.7 1.2
2022 41.7 5.1
2023 51.3 6.5

2023 being an exception, since the 10-K isn’t out yet, and these numbers are approximate.



I don’t have a position in AEHR, and instead merely taken an aercheological interest in the company.

33 Likes

Thanks, @Raylight, this last graph you posted,

is really helpful for getting a sense of what might be ahead for Aehr as other customers move from using their test systems for product engineering to using them for production. The volumes go up significantly as the production lines are scaled up.

Note that once a customer has moved to the point of using Aehr’s systems in production, the competitive risk is greatly reduced. Selecting a test system to use is a huge deal, involving many different groups in the company. It’s not a decision that’s readily changed.

However, until a customer reaches that point, a competitor could still take the account away. The customer could still be evaluating which system is the preferred one – so Aehr is clearly in the running, but they could still lose out.

I haven’t yet bought any Aehr, but this conversation is making me much more interested in taking a “starter” position, so I can learn more.

ActonUp

25 Likes

Found this today, could be a new use case for SiC - this is just a few units but still interesting.

  • Meta Platforms may only build roughly 1,000 units of a new pair of augmented reality glasses slated to be released next year.

  • The U.S. tech giant will reportedly use the units for internal development and public demonstration, The Information reported.

  • The glasses will incorporate silicon carbide in the lenses and a complex assembly arrangement, with factories in China, Taiwan and the U.S., the news outlet added.

Hannes

14 Likes

I doubt that has anything to do with Aehr. Aehr makes test equipment for SiC applications in semiconductors, primarily those use in power inverters. SiC used as a lens material (or whatever) is not likely to require the kind of testing for which Aehr sells equipment.

12 Likes