And there it is, SMCI’s first to market differentiation at play:
Nvidia said Super Micro and Quanta will be the first 2 companies building servers based on Nvidia MGX server specs, which help builders cut cost and development time building over 100 server variations for AI, HPC and Omniverse, media report. Shipments to start in late August.
DigiTimes reports that the price of an AI server is 3 times higher than a regular server.
Hi Rob,
Thanks for interpreting the recent news on SMCI. Can you help me understand what numbers I might look at to make some guestimates about EPS and PE? I’d like to take a position in SMCI with appropriate risk management, but I don’t know what stop losses I’d consider after such a parabolic run-up. I’ve got the ‘feels of UPST’ tingling and as I’ve exited NVDA for similar valuation reasons, I’m struggling to put together a cohesive plan on building positions in these non-SaaS companies with exposure to AI and GPU fabrication and utilization.
Nobody wants to see where I pulled those numbers from!!!
As a hint… the sun doesn’t shine there…
As I said… guesses. But based on extrapolations from current performance on earnings and using a conservative view of reasonable PE based on PE being close to growth rate.
As for stop losses… most folks at the Fool would have one word for that: don’t.
Rob
He is no fool who gives what he cannot keep to gain what he cannot lose.
Rob,
I’m very happy for you 2023 results have been good, but this isn’t really a portfolio, it’s a gamble on mostly one company. I hope it works out, but please refrain from posting it on Saul’s board.
TO ALL: Rob is far from the only one who has done this lately – this message is to everyone:
Please do not post trades here on Saul’s board. This board is about discussing individual companies. Some people share their portfolios monthly and discuss the companies they hold, but hopefully it is clear that this is very different than posting about a trade you’ve made, or stating that you are allocating 50% or 70% or 90% to XYZ. There are other boards and sites for trading. Here, please keep the focus on the companies.
Thanks,
Bear
Assistant Board Manager
I will comply. Sorry… got carried away… 
That’s why this is the Motley Fool… different views are welcome and encouraged. Most of the time…
Rob
He is no fool who gives what he cannot keep to gain what he cannot lose.
Sorry Rob, but I’ll have to delete that post.
We have a responsibility to all the thousands of viewers who look to this board for sensible investing ideas, but talking about having 70% of your portfolio in a very speculative stock that has already had a huge run-up, and goosing it with options, is really not what this board is for. It’s for analyzing individual companies. (Discussions of individual option contracts are off topic for our board anyway).
And since you have acknowledged losing 94% of your total portfolio in the recent sell-off because of using options should give you an idea why they are OT for our board.
Saul
THIS is interesting!
From the Founder/CEO’s keynote address at Taiwan’s Computex conference:
"We are expanding our solution offerings with innovative servers that use the NVIDIA Grace CPU Superchip and are working closely with NVIDIA to bring energy-efficient servers to market for AI and other industries. Worldwide our manufacturing capacity is 4,000 racks today and more than 5,000 later this year.“” (emphasis added)
There was more, relating to the customization of their servers, as well as the energy cost savings due to liquid cooling and who knows what. What is interesting is the bold part because up until NOW… SMCI has been telling investors that they were planning to have capacity of 4000 racks (per month)… by the END of 2023.
Now… they’re saying they have 4000 capacity NOW… with 5000 by end of year.
Very positive news… and in my view, a nice aggressive stance by the company.
Also… “assuming” they are already producing 4000 racks NOW, that means their earnings guidance is too low… and their upcoming guidance will be more aggressive than I anticipated.
So… good news all around, IMO.
Rob
He is no fool who gives what he cannot keep to gain what he cannot lose.
Great find @XMFRob !
Per their latest earnings call, they said (requoting from my initial post in this thread)
Having high power efficiency and air/liquid thermal expertise has become one of our key differentiators of success.
Provided there are no supply constraints, we can design, build, validate full systems and deliver turn-key rack level solutions to customers within a few weeks of placing an order instead of months from competitors …we are on track to support up to 4,000 racks per month of global manufacturing capability and capacity by the calendar year end. We are doing so by efficiently taking market share in the new and fastest growing markets.
That is VERY incredible and positive if what they claim is true. “4000 today”….and “more than 5,000 later this year”
Significantly raises the chance of a massive beat and raise on their next earnings.
Note: The last earnings call with that “4000 racks by the end of the year” was uttered on May 2… so this is a big upward change in just four weeks.
And, IMO, it’s significant that this is a Founder-led company with the Founder having “just” bachelor and masters degrees in electrical engineering… started the company 30 years ago in 1993. SMCI has ALSO been changing their company target from $10B in revenue (which is STILL a target)… but their emphasis is shifting toward reaching $20B in the “medium term” (undefined). They’re aggressive and they have their sights set high.
Rob
He is no fool who gives what he cannot keep to gain what he cannot lose.
Charles and Jensen presented SMCI latest products and while it was difficult to understand Charles, it was great seeing the very good chemistry between these two guys and obviously companies.
It was very good seeing the passion about their products and Jensen did a great job on actually selling the new products from SMCI.
Charles is definitely more the techie nerd guy and very passionate about progress and developing new and faster products.
They were especially thrilled about the new era driven by AI coming up and I am very happy, that I found SMCI through this board and will hold for the foreseeable future.
I am quit surprised, that - although the stock price moved up tremendously in 2023 - I don’t see many people adding SMCI to their portfolios. The stock is still reasonably priced and the growth opportunities ahead are enormous.
Best
Haikilikona1
I will not pretend to know much about SMCI, but I can speak to some of the reservations some have, as I mostly share them.
My prior beliefs (before SMCI early-May earnings report):
SMCI makes hardware, a commodity business, and though they’re profitable they have very low margins and it’s not super easy to grow and scale. Therefore they have a PE ratio of ~10, which makes sense to me.
What changed? (my casual spectator version – feel free to correct me):
SMCI guided for a big bounce back next quarter (after a disappointing one this past quarter). But the guide is for only 16% more revenue than the quarter one year ago. Then Nvidia guided for a big quarter. SMCI will benefit from the same demand…so maybe people think they’ll do a lot better than +16% this coming quarter, and for many quarters thereafter. (Note: a lot of the reason people think this for SMCI and NVDA has to do with expected AI spend. There’s a lot of optimism around that right now, and the market has bid things up so the optimism is getting expensive.) SMCI even increased their capacity goal from 4000 to 5000 somethings.
My updated beliefs:
Ok, maybe that kind of growth is enough to justify a PE higher than ~10. But it’s more than doubled in a month. I don’t think that makes it “reasonably priced” just because it’s cheaper than tech stocks or recurring revenue businesses. But maybe the demand will increase way more than I think. I just don’t know.
There you have it. A lot of us see a business that we didn’t know much about or think much of that is obviously seeing great demand right now. We don’t know what to think about it still…and we certainly don’t want to buy after the price has more than doubled…unless we can get our heads around it.
Bear
I am trying to get a gauge of SMCI vs. its main competitors although they each have dif product/platform offering.
Below numbers are a quick pull from Seeking Alpha. Note Rev growth is YoY forward projection and P/E is non GAAP forward
Dell: Rev growth -3.4%, P/E 8.5
Cisco: Rev growth 5.5%, P/E 13
HPE: Rev growth 2.6%, P/E 6.9
SMCI: Rev growth 30%, P/E 21
We sure like it when SMCI P/E is 10 when it is in a better alignment w/ the peers. However, its growth rate propels its PE to a higher level.
SMCI grew its revenue from $3.6B in 2021 to $5.2B in 2022, that’s 46% YoY growth in 2022.
Then for FY 23, its rev is projected to be $6.7B (with only 1 qtr to go), so that’s 30% YoY growth in 2023.
For 2024, they are projecting at least 20% growth announced at May earnings, with 4K rack/mon. Then they updated to 5K rack 2 weeks later. So that’s about 25% upside.
Assuming 25% YoY sales growth translate to roughly the same 25% earnings growth, then next year’s earnings will be $10.75 (2023 earning) x 1.25 = $13.4/share
Assuming P/E close to its growth rate of 25, then the stock is prob worth $13.4 x 25 = $335. If given 25% margin of safety, that brings us to $335 x 75%= $250
Still some room to run.
Would love to hear all of you guys thinking. Much appreciated
However, I do have some reservation regarding the management projection.
For example, at Q2 23 (Jan 23) earnings call, CEO projected 30 to mid 30% rev growth for the year. When pressed by an analyst, he decided on mid 30% growth. Now based on Q3 earings, it is more than likely 30%
Two, at Q2 earings call, they commented that “Our U.S. facility still has 40% capacity while Taiwan still has 50% capacity headroom to grow for the next one to two years”. Yet at Q3 earnings, they explained the short fall of rev as supply chain shortage of key components, mainly due to surging AI chip demand. It gives me the feeling that even the management did not quite anticipate the AI surge thus did not deploy the key component production/shipment properly. It makes me wonder how solid is their “20 to over 50% YoY growth for the coming years (at Q2 earnings call)”
Except none of the ones you listed actually compete with SMCI.
** To understand how Inspur and SMCI are similar to each other and different from Dell (DELL) and Hewlett Packard Enterprise (HPE), it can be helpful to understand what Lenovo (OTCPK:LNVGY) describes as ODM+. ODM+ is targeted at hyper-scale and/or very large and sophisticated enterprises that have custom needs for their own data centers, and reduced need for support.
The largest players in this space are, in order of size, Inspur, Super Micro, and Lenovo, and Inspur has now been largely removed from the picture, and while Lenovo is still a player, the fact that it is headquartered in Beijing cannot be a positive in the current environment. Supporting the bull case for SMCI, Lenovo was very bullish on ODM+ on their recent call.**
Also
** Inspur largely relies on US chip makers like AMD (AMD), Intel (INTC), and NVIDA to supply the chips for their servers, and being on the entity list will limit their ability to meet global demand for server hardware. This ban, especially if it is actually as restrictive as Huawei (i.e., not just advanced chips but most chips), will limit Inspur’s ability to do business anywhere outside of China. At a recent conference, Nvidia management stated:**
Andy
Hey Bear,
I definitely see your points, and just wanted to say that I see a lot of parallels between SMCI and AEHR, which several on the board have waded into.
Both are in the hardware business, which could be commodified or have a technical advantage (NVDA for example technically makes hardware in the sense they manufacture cards, but there is a seemingly very large technical moat there) - AEHR seems to have an advantage here in the sense that their packaged burn-in process isn’t replicated across the industry, but hard to say how much of an advantage this actually is, can the increase in thoroughput be replicated easily, without violating any patents? I’m imagining putting a bunch of chips side by side for testing, but surely it’s more complicated than that. SMCI doesn’t have a technical advantage, but rather a relationship one with NVDA. The thought here isn’t necessarily that they have the best technical product, but that their products have the best technical product in it. That the NVDA CEO is partnering and appearing in panels lately with SMCI seems to be the number one thesis for this company, at least to me.
Valuation is a concern for SMCI, at 20 PE, but I don’t agree that 10 PE seems appropriate because it’s a hardware business. We’re used to evaluating P/S here, and thus margins really matter when surmising whether a stock is overextended. But P/E already takes margin into consideration. What matters when considering whether P/E is cheap or not is forward growth. SMCI can report only a 10% increase for the coming Q2 quarter in EPS, but if on the call they indicate that it is expected to double, the market will react to that piece of news, not the actual results (NVDA’s Q1 results was a 13% drop from prior year Q1, but stock went up 30% because they indicated the next Q they are expecting a 57% QoQ growth, and this is after it had already doubled from the lows in Nov 2022).
Valuation for AEHR seems to me is a much more dangerous looking number, at 85 trailing PE, and afaict that’s based on an expectation that the 2 new contracts in the pipeline and the newly announced contract will add significantly to their bottom line. However, it’s hard to tell the impact since there has been no concrete numbers. Feels like an overload of optimism for both stocks when it comes to valuation. I will say that it’s easier to grow from $17m (AEHR last Q) than from $1,283m (SMCI last Q) - but then again SMCI has much larger, enterprise customers
And speaking of customers, the biggest reason I haven’t added AEHR but have added SMCI is because the concentration is scary. 94% of AEHR revenues are from two customers, of which 82% is from the top customer. (not a typo) Saul has indicated that he feels AEHR has the power here because of patents and pressure from customers downstream on Onsemi, but I see it differently. If your sole bread provider threatens to pull its business, even if you know it’s an outright bluff, you’re not able to fight back and avoid decreasing prices, because if they do follow through with the bluff, you’re dead. It could even be non-AEHR related, for example, if Onsemi loses a bunch of its own contracts with its customers? Or if some Onsemi chips fail and they need a scapegoat? SMCI has issues of its own with the reliance on NVDA, but they can at least source chips from other players in the industry, so with concentration, it’s a supplier issue with SMCI but a customer issue with AEHR. Hopefully the new contracts AEHR have talked up will be large and repeat customers, but the future to me seems harder to tell with AEHR rather than with SMCI.
Hey @Aphalite – I’m not seeing the parallels you are between SMCI and AEHR.
A business like AEHR with gross margin passing 50% recently…that’s just very different from a business with a sub-20% gross margin. AEHR isn’t a commodity.
You make a very good point here:
(And I would specify that revenue is the growth to focus on. Not sure how much more operating leverage either of these companies can squeeze out. Maybe a little, but revenue will be the main driver of earnings growth.)
Forward growth is the reason I believe AEHR is not so crazy expensive as it might seem at $41 with a trailing PE of 68 (not 85). I think they could grow revenue at hyper rates (I’m hoping for well over 50% YoY and maybe much more), and not just next year, but for many years. This is very much a speculation at this point, which is why I am keeping the position really small.
Similarly, if you believe SMCI can grow revenue at even something like 25-30% for the next several years, it probably makes sense at a PE of 20. I have no idea how likely that is at their scale, but if they guide to much less than that at some point, I predict the PE will come down from 20. But again, I don’t follow the company…I’m just telling you what I think is expected at this price.
Bear
I look at AEHR more as a small ASML. Growth for years with a moat around their testing patents. As far as SMCI it is going to be very cyclical and is in a multi year growth with the data centers being updated but is trading past the top of it’s historical norms.
Andy
Rosenblatt Starts at Buy, ‘TAM is expected to exceed $178 billion by 2027’ PT $300