Well, if you’re going to talk about the price of oil, then you need to also talk about all the government incentives to adopt renewables. At present, with brand new war in the Middle East, oil prices are rising. So with oil going up and incentives aplenty (with solar about to expire unless Biden extends it), the demand for solar and BEVs should be high.
I think Musk nailed it. The number on concern when buying a new car is the payment. So long as interest rates are high, the monthly payment will outweigh every other factor.
On Saul’s board forward business momentum counts. I remember when Saul got out of Zoom because sequential growth was on 4% while year over year growth was substantial. Saul was correct. Of course nothing is 100% but it is a good process.
With AEHR it is caught in this sudden slowdown in EV demand by potential high volume customers. Most of these OEMS are not selling SiC based EVs. It is their new models that are. Thus why when they slow down or delay production of their new SiC enabled vehicles it is a big issue to suppliers.
The slowdown is real. At some point it will reverse. One big new customer announcement could revitalize AEHR. But if customers keep holding back because of uncertainty with their own OEM customers then you have the opposite issue.
The thing about these troughs is they eventually recover. Stocks crashing because of this and recovering big time down the road is not uncommon. SaaS for example had a big crash in 2015.
The method on this board however is to follow the business momentum as that is far more likely to “speak truth”. Here we may just have the trough part of new technological platform. We don’t know but it is likely and even if it is we don’t know for how long and how low it may go.
So it makes perfect sense to sell off AEHR using more of the “Saul” method (which is quite effective) and perfect sense to buy for long term recovery (which, and Saul has proven this not to be the best strategy usually - I particularly remember when he compared the math with NET’s more technology superior competitor (name I cannot even remember) and how devastatingly superior it is to hold the winner as they climb the S curve instead of trying to play turnaround.
It is possible both methods are correct but over different time frames. But it is easy to see why there is present unease with AEHR as EVs are definitely having a rough patch here.
SiC designs will continue even if new models are delayed. The big 3 and others are frantic to narrow the efficiency gap ( and cost gap) with Tesla models. They must introduce improvements along the way in existing models. Just like Tesla has continuously done.
Where is this “slowdown in EV demand?” Not in the US:
EV sales surged 49.8% compared to last year…
The data shows most automakers have picked up the pace significantly over the past year.
Overseas…Mercedes-Benz EV sales climbed 284% to reach 10,423 in Q3.
It’s why with Tesla alone growing sales 38% YoY, it’s share of the so-called “EV market” is decreasing (38% is less than 49.8%).
Now, it’s true that not all of these EVs, even in the US, are SiC-based, but that’s clearly the future for EVs, and EVs are clearly the future for personal transportation, at least until robotaxis are real.
For me, the questions about AEHR are around what is the real TAM for their $2.5m machines? If anything, their machines are so capable that it may not take that many of them to satisfy demand, and the additional design work for each new SiC chip may also be slowing down as automakers and Tier1s settle down on SiC chip design improvements.
After that, I ended up investing in AEHR, but at least two of the three questions I asked in that June post remained unanswered. I decided to invest largely because of the huge market opportunity.
Now, it looks like customer concentration has become an issue, but it’s always been there. Also, not sure this has been addressed somewhere, but I’m going to bring it up again: they say they have an installed base of 2500 machines; where are they? I wonder whether anyone has clarity of that.
In any case, I still have my AEHR position and still believe in the company and the market opportunity.
Tesla sales growth is slowing. Since Tesla is the end market for 88% or more of the chips tested on AEHR equipment that means this portion of the revenue stream is likely to slow.
Ford is indicating its eF150 sales are not good, having cut a shift from the plant and delaying their new battery factory. GM is delaying the building of its new all EV plant and delaying production of its next generation pickup trucks for multiple different trucks.
Those are SiC enabled products. They are having their production slowed or delayed at least a year and new plants long planned delayed (at best) at least a year.
This is not limited to the US either. VW is doing the same sort of things with their EVs.
Thus, a slowdown. Particularly a slow down in next generation SiC based EVs that had been long planned and Musk talking about the fetter of high interest rates.
A saving grace may be the new Tesla cybertruck that uses 800volt battery and ON and STM will need to produce SiC chips for this.
Some of this delay may be associated with the strike that is going on. Some with fear of higher interest rates as Musk articulated. Some just based on less consumer interest in the vehicles they planned at present.
Lower interest rates and resolving the UAW strike will go a long way to resolving 2 of the 3 issues anyways. So EVs are far from dead. Trough of disillusionment is Gartner Hype Cycle Research Methodology | Gartner.
Doesn’t mean it’s dead does mean original estimates are being reconsidered and optimistic plans from lessor players being re-assessed.
An electric passenger SUV is great. An electric work pick up less so. Eg, Rivian has an awesome new truck out. Car and Driver loved it. Problem with it. Its 400 mile range drops to 110 miles when towing a boat. Tough to drive your boat to lake for vacation if you have to recharge every 100 miles.
The utility and value points for EVs have to be better ascertained. Tesla does this. Their semi truck is made for in-town deliveries. To return to base every night for a recharge. It is not meant for long haul. Tesla will build a semi for long haul as well but not its first model.
Also, Ford, BMW, et al will be joining Tesla’s recharging network in 2025. Another utility factor for EVs is access to Tesla’s fast charging network.
I am not an executive in the industry but these are things that are unfortunately visible to anyone interested in the topic.
Ford would not be cutting a shift from its core vehicle the F150 electric unless these demand issues exist from a utility value perspective. Tesla can make an EV to fit its purpose. Ford and GM seem to just want to put batteries in their existing trucks and sell them at high prices without considering the different elements of gas vs battery platforms.
Anyways, lower interest rates and resolving strike will help. But so will reassessing what EVs do best and building products to address these strengths that customers will want at price points they can afford.
Even Musk specified the delay in their Mexican factory is due to worries about high interest rates and world economy. Just being paranoid and cautious was basically his explanation.
Where are you getting this from? Aehr’s 88% customer concentration is Onsemi, which is one of two sources for Tesla’s chips. Are you saying that all of Onsemi’s business with Aehr is for Tesla?
I am saying the probability is high that most or all of it is testing chips delivered to Tesla. Find out who else ON or STM has as customers needing SiC wafer tested product for 12 hours or more of testing. Let us know if you find someone and how large their chip needs are compared to Tesla.
There are some other considerations regarding SiC. Here are a few things to think about:
Coherent Technologies (COHR) announced on Oct 10 a joint venture with Mitsubishi and DENSO. Those two companies will each invest $500M for a total of $1B contribution invested in SiC manufacture of 150mm and 200mm SiC wafers. They will each receive 12.5% of the business for their investments while COHR will retain 75%. COHR will operate and manage the business as a separate subsidiary. Mitsubishi and DENSO have entered into long term supply agreements with COHR for SiC wafers (Coherent’s Silicon Carbide Semiconductor Business to Receive $1 Billion in Investments From Denso and Mitsubishi Electric | Seeking Alpha).
WOLF, a global leader in the supply of SiC will report earnings on Oct 30 after close of market. I can’t predict what they will report, but no matter what they say, it should shed some light on their SiC future plans. IMO this will be material to AEHR’s future as well.
Infinnion (IFNNY) reported during 3Q23 quarterly CC that they substantially increasing their investment in SiC in order to maintain their market share in this fast growing market segment. They are estimated to report 4Q24 no 11/13. I assume their report will shed more light on SiC production.
ON will report 4Q23 on 10/30. As Aehr’s largest customer, their report should be very informative with regard to when Aehr might realize the much anticipated growth.
STM will report 3Q23 before open of the European market (that’s a stupid early hour for us US investors). As we know, STM is heavily involved in the production of SiC products. I have not researched their customer base, but I assume they are the primary supplier of German and other European car makers.
In short, I think we will get a lot more visibility about Aehr in the next few weeks. I’ve not sold my shares at this time. Of course, if these companies reports a slow down in SiC production, Aehr’s stock will likely take yet another hit. Hello $25/share or less. OTOH, it could go in the other direction.
Tesla isn’t the only auto company ON is supplying. Here is an article from earlier this year on BMW entering a long-term supply agreement with ONsemi to provide them SiC parts for their next-gen EV’s.
Further another article of BMW’s EV unit showing strong growth. 100%+ growth in deliveries in Q3 and expected to remain strong. BMW has ~4.2% share of the EV market in Q3 so it is not small.
Let me add I feel like there is a lot of analysis by price going on with AEHR. They give guidance for 54% growth and $100m+ rev this year during there Q4 report and the stock pops 20% to almost $50 and everyone is happy. Now after this report where they maintain the exact same guide that everyone thought was amazing 3 months ago and people are disappointed. I don’t see any major changes to the outlook between the 2 Q’s besides people looking for a reason the stock is down.
That is for BMW’s next generation cars. Just like with GM and multiple others the SiC models are coming out in 24 25 ‘26.
I assume ON will need more from AEHR for BMW when these new models start being produced in volume. So that will be a good start diversifying away from Tesla as the end company.
The (relevant) installed base is nowhere near that number. I tried to address your question a while back, but the reply ended up in another thread:
Excluding ON, the total revenue for wafer-level products in the last decade (40 quarters) is somewhere around $105 million. This number is an estimate, and includes contactors and services.
By 2008 the installed base of the FOX product line was around 125 systems, of which I presume the majority was FOX-1 (single wafer/slot). I’m not aware of any more recent disclosures, but the total revenue puts a limit on how many systems they’ve sold since then.
The language used is somewhat misleading in my opinion. But it’s now removed from the latest 10-Q/10-K. Unfortunately, so is the part about naming their >10% customers, which means less transparency and more unknowns with regards to customer concentration from now on.
Should note that it’s easy to underestimate the throughput of these systems. E.g. at 12 hours per cycle, ~75 FOX-XP systems would give you a capacity of one million wafers per year. A move from 150 mm to 200 mm wafers would reduce that to ~40 systems to cover the same demand.
The FOX systems are built to handle 300mm wafers. It seems that the largest wafer yet possible for SiC and GaN is 200mm. However, Si wafer production is moving to 300mm.
So the 300mm capacity of the FOX systems is forward looking. Either looking forward to larger SiC and GaN wafers, which at present is not possible, they are both just difficult to manufacture - the material is brittle and tends to fracture with larger wafer size. Or forward looking to that elusive memory market? Either way, the FOX machines are already built to handle the 300mm wafer, should an application develop.
Aehr doesn’t name the top customers (did they ever?), but the customer concentration disclosure is certainly still included in the 10k (I think it’s required). From page 7:
During fiscal 2023, two customers accounted for approximately 79% and 10% of the Company’s net sales. During fiscal 2022, one customer accounted for approximately 82% of the Company’s net sales. During fiscal 2021, four customers accounted for approximately 24%, 23%, 20% and 10%, respectively, of the Company’s net sales. https://www.sec.gov/ix?doc=/Archives/edgar/data/1040470/000165495423011271/aehr_10k.htm
Even if they get more 10%+ customers this year, it’s probably safe to believe OnSemi is still driving this train way more than any other customer.
Aehr has named the top customers on an annual basis since IPO. From last year’s 10-K:
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.
You can derive a lot insight from this information, like I did in the post linked above. For example, that AEHR recognized 16.7 million in revenue from STMicroelectronics during fiscal 2017-2020, anywhere between zero and 1.1 million in 2021, and anywhere between zero and 5 million in 2022. Or that Wolfspeed has never been a 10%+ customer, at least up until fiscal 2022.
In any case, should have made it clear that there’s nothing suspicious about it. I was thinking about the effect of having less information available.