SMCI--Scratching My Head, but Buying Like Mad

So SMCI preannounced earnings yesterday, telling us that the new net income guidance for 4Q, to be announced August 8, will be at least 23% higher (corrected–see discussion below) than the guide they gave us May 2, when they were already extraordinarily bullish.

Here is what the CEO said on May 2:

“Supermicro continues to see record levels of engagements in our new generation product lines, especially for AI applications," said Charles Liang, President and CEO of Supermicro. "We secured several new and large design wins and are deploying some of the world’s most leading GPU clusters. With the recent new key components supply chain challenges mostly in the rear-view mirror and production normalizing, we expect to gain share and expand scale as we emerge as the true leader for rack-scale Total IT Solutions.”

So they’ll be at least 23% (corrected–see discussion below) better than that.

And how does the market respond today? By treating SMCI like just another stock on the Nasdaq. It’s up less than 1% as I write. I upped my stake from 2% yesterday to around 13% today.


The stock ran up quite a lot in the past year (≈550%). So very, very good numbers were widely expected - especially since $SMCI is tied closely to GPU demand and NVIDIA had a huge quarter recently.

How did you calculate the 16%? Comparing the higher end of the previous guidance to the lower end of the pre-announced earnings, I’m coming up with a figure closer to 23%. Am I missing something?

While I have a position in the company, I don’t think this is as much of a no brainer as it might appear. $SMCI remains a hardware company with low margins in a highly competitive and largely commoditized sector. The company’s earlier relative undervaluation has more than dissipated, and it has a higher valuation than its competitors. This could potentially mirror Zoom’s trajectory, experiencing a period of rapid growth due to AI advancements, but struggling to sustain that momentum.

Just a word of caution because if something seems to good to be true, its even more important to not get to greedy.



Sorry–you’re right. 23%! I take your point about Zoom and about the margins. I don’t think it’s an apples to apples comparison, but I confess I don’t have enough information to say that with confidence. On the other hand, several analysts raised their price targets following this news and presumably they have more information than I. (Then again, probably several analysts raised their PTs for Zoom at its peak.) But on the other other hand, tight margins and massive share price appreciation notwithstanding, we’re still talking about a company that is trading at like 30 times trailing earnings versus like 230x for NVDA.

Regarding the price action, it looks like SMCI released this information before the open yesterday, so it may explain why SMCI fared better than most stocks yesterday. Still, I think the news got buried with the Tesla earnings and the overall market pullback that ensued, and also by the report from Taiwan Semiconductor Manufacturing (TSM), though I have seen some commentary on Twitter (which I tried to verify by reading through the TSM transcript) stating that some of TSM’s difficulties bode well for SMCI.

I believe the relevant passage from the TSM call may be this one

For the AI, right now, we see a very strong demand, yes. For the Tongan [ph] part, we don’t have any problem to support. But for the back end, the advanced packaging side, especially for the cohorts, we do have some very tight capacity to - very hard to fulfill 100% of what customer needed.

I believe advanced packaging, whatever the heck that is, is what SMCI does.

So to sum it up, I think that when SMCI releases earnings August 8, or maybe as investors turn their full attention to SMCI in preparation for August 8, this preannouncement will catch the market by surprise. Sounds weird, maybe, but I have seen far weirder things happen.


Yep, $TSMC headline results certainly were a factor. But I also see their comments as a positive.

Much like you, I don’t consider them overvalued, but I no longer see them as undervalued either. I believe the demand for GPUs, and therefore building block solutions, should persist for the foreseeable future.

The most significant risk I see is the intense competition within the industry from firms such as Cisco, Dell, HP, Lenovo, etc. To maintain their position they have to always be first with new technologies, while also remaining cost-effective. There are many alternatives out there, all of which will be doing everything they can to gain ground.

Important to keep a close eye on their margins going forward, as well as any hints of negative shifts in market share.

Edit: Because you mentioned advanced packaging - supermicro has nothing to do with it. This is a technology semiconductor companies use to assemble chips in a more efficient way. $SMCI builds server solutions - basically the stuff where the chips sit inside.


“Packaging” in semiconductors is placing the small slab of silicon into the plastic or ceramic integrated circuit, then bringing the electrical signals out to the metal connectors that connect the insides of the integrated circuit to the printed circuit board.

SMCI builds high speed computing servers: Like your PC, but made to fit in racks hundreds at a time for so-called “server farms”. Their product includes printed circuit boards upon which the integrated circuits are mounted.


I‘m still trying to figure out what’s going to happen to SMCI. It‘s my biggest holding at the moment and I bought heavily in the 2nd half of May. Therefore my position is up a whooping 60% and I recently took some profit off the table, because the holding grew way too large in my portfolio.

I‘m still convinced SMCI is a great pick & shovel AI play for the next couple of years and I will keep a full position as long as the numbers move into the right direction. The pre-earnings report definitely was great and did beat earnings consensus on revenue and EPS.

SMCI was way undervalued, compared to it‘s peers, but closed that gap over the last 12 months. Now, thanks to the AI hype, it overtook it‘s peers, but underpinned by a great pre-earnings report, valuation on a NTM is back in line. So from that point of view, there should still be more room to grow.

Keep in mind, that SMCI has a market share of 4% and Jenson Huang sees a 1 trillion data center equipment overhaul over the next couple of years, that could lead to roughly 40 billion in revenue for SMCI over the next couple of years.

Market Share SMCI

1 trillion overhaul

But what still bothers me a bit is, that SMCI is still kind of flying under the radar. There are still only 6-7 analysts covering SMCI, only 3 analysts from smaller companies attended the last earnings call back in May and press coverage is negligible, too.

So that’s leading me to my first question: What’s going on with SMCI?



This is one I’ve been watching. The pre-announce was a huge beat, but the reaction suggests most if not all of that had been priced in. I wouldn’t be surprised if many holders decided to harvest gains until the actual Q4 report with the initial FY24 guide. That should give us a better indication of just how long management thinks this surge will last. I’ll reassess after that.

I did see a comment on tech cycles recently which makes a ton of sense. When new tech is unveiled, the order of big winners goes picks & shovels → hardware → software. Clearly we’re all scrambling to find those with AI. NVDA’s chips look like the first obvious picks & shovels win. With SMCI’s servers so tightly tied to those chips, SMCI seems to be sitting on the front end of that arrow into hardware. I’m also curious to hear if there is enough pricing power for any new server offerings to bump margins at all. That’s the lens through which I’ll be viewing SMCI’s guide and comments. I have no idea what that FY number should be though. I’m guessing the market will let us know that.


I view that rationale as entirely irrelevant. Why?

  • Their margins… regardless of the competition… remain constant to slightly rising.

  • Their earnings… are their earnings. It matters not what the margins are as long as they are not dropping.

  • Growth in earnings are… and have been… high. High growth in earnings is typically rewarded with a PE that is similar to that growth AT LEAST. That’s the PEG. And it often results in a PE higher than the growth rate. While the market may not give it… the growth in earnings arguably deserves a PE of at least 30.

  • Earnings for the fiscal year ending next June are likely to be at least $15 a share. Regardless of what anybody thinks about poor margins. That suggest an opportunity for a 50% share price increase.

Will it happen? That’s the future and that can’t be forecasted for certain… but the company performance supports it.

He is no fool who gives what he cannot keep to gain what he cannot lose.


After they reported the 2022 fiscal year on 8/9/2022, and had $5.65 eps (up 128% YoY), shares traded around $60-70. At the time they also guided for $7.50 eps in fiscal 2023. So that’s about a 33% growth guide. But the fwd PE was only ~10. Why do you think that was?

…but they’re going to do ~$11.75 eps, good for 108% growth…either the 33% was a colossal sand-bag or it’s been an unexpectedly great year. Even if they guide for the $15 you’re calling for in fiscal 2024, that would be a 28% growth guide. If the fwd PE were ~10 again, the shares would only command $150. But currently shares are sitting at ~$300 which means the fwd PE is about 20.

So growth is expected to be 28% vs 33% a year ago…but the PE is double. The market has re-rated SMCI’s multiple up by 100% in a year’s time. …for now. But why do you think it started at a fwd PE of 10 a year ago, even though expected growth then was higher than it is now? And what gives you confidence they “deserve” not only this multiple, but an even higher one?



I think Nvidia’s last earnings report changed everything.


Something that stands out from these numbers to me is sales beat the guidance by 20.3% but income beat by 38%.

It seems like this would imply the profitability or margins are expanding on selling the devices? Or is there something else to account for why the beat on income is roughly double the beat on sales?


It is very lumpy quarter to quarter and the only way to make sense of it, is to look out YoY. Quarter to Quarter it is hard to tell what might happen although with the Data Center expansion this could be a wave moment going forward, But how long will that last? Right now it is trading at the Top of it’s P/E but way above it’s P/S. I think every quarter you are going to have to listen and see where it’s heading and expect that if some bad information on the future come out it is going to tank.



Regardless of what you may think the expected growth may have been before is irrelevant.

The current growth is very high and that is building on the high EPS. And PE is typically a function of growth. Like it. Don’t like it. Think growth will be high… or don’t. We shall see.

Of course… you’ve said… I don’t have a portfolio and SMCI is part of this non-portfolio. So maybe my thoughts are irrelevant as well.

He is no fool who gives what he cannot keep to gain what he cannot lose.


Bear I suspect you’re just being a devil’s advocate here, but I’ll bite. SMCI is in a cyclical industry. In boom times the market expectation is stronger EPS growth. In down times the market expectation is a shrinking business. If you look at historical PE for SMCI (quick and dirty version here), it has bounced anywhere between the high single digits to the low 30s for the past 13 years. And you can see how its EPS graph has looked like a wave, cresting up and down.

Currently the market is expecting long tail multiple years of continued growth because of Nvidia’s May earnings and disclosures. Will it hold? Depends on how long this datacenter buildout lasts. Imo not a set it and forget it holding (although none of the stocks we discuss on this board are). So you are right that the multiple will eventually be rerated downwards, if growth slows or reverses. But isn’t that the case for all of our stocks? Even those of us who’ve gotten in “late” are sitting on 30%+ gains currently. Personally I think the trend hasn’t broken yet, so I’m continuing to hold. Valuation wise it’s still a better play than Nvidia, although the main benefits of this AI craze will seemingly accrue to Nvidia at the end of the day.


I try to stay away from cyclicals, commodities, story stocks, chasing big price moves (like trying to figure out whether as you put it the “trend has broken”), and really most companies without recurring revenue.

You’re of course not me, so maybe you have different feelings about all that, which is your prerogative. I do appreciate you clarifying that a cyclical (and a somewhat commoditized one) is what this is.

That said, I surely don’t know where SMCI will top out…I hope it goes to $400+ for y’all. But I’ll stick with companies I feel can have more revenue year after year. Those are difficult enough for me.



Alphabet and Microsoft reported earnings yesterday. Both pointed out, that generative AI will be further driving their revenue and therefore started and will continue investing to further build their data centers to keep up with AI demand. That should be very good news for SMCI.

Quotes are from todays FT.

Both companies pointed to higher spending in the coming quarters as they build out the data centres needed to support an expected boom in demand for new generative AI services. Facing a barrage of questions from Wall Street, however, executives at both companies stopped short of putting a figure on the capital spending binge to come, or predicting how strong the revenue uplift from AI might be.

Chief financial officer Amy Hood said that a revenue lift from new AI services such as the “co-pilot“ being added to Office software, would not show through until after the end of this year. She also said that capital spending would rise each quarter over the next year as the software company seeks to match its higher spending to the pick-up in revenue growth from AI.

Ruth Porat, Alphabet’s CFO, said the search company expected “elevated levels of investment in our technical infrastructure, and that would be increasing through the back half of 2023 . . . then continuing to grow into 2024”

Long SMCI.



Can any of the SMCI bulls on here explain what the barrier to entry is like for competitors trying to steal SMCI’s GPU server business?

Designing a server on a printed circuit board and building it is a commodity business with CPUs. SMCI has built a great new business line doing the same with GPUs.

Hooray, I’m happy for everyone who has been a shareholder and participated in the stock rise – but what’s going to keep SMCI out in front as the competitors pile in?

Initially I expect NVDA to be allocating their scarce GPU chips to their best customers, and SMCI probably qualifies. Over time, however, as NVDA scales up their manufacturing volumes, how can SMCI retain their edge?



The barrier to entry is the relationship. SMCI was the first Nvidia partner for data center GPUs. Unfortunately relationships are a very squishy factor which you can’t really measure, since it can evaporate over night like you’re alluding to. Nothing is preventing NVDA from going out and getting a better deal in the future.

On this point I equate SMCI a lot to AEHR - whereas AEHR has a potentially catastrophic event with customer concentration, SMCI has a potentially catastrophic event with supplier concentration. Although revenue wise NVDA products accounts for 20-25% of products currently, that’s where most if not all of expected forward growth will come from. Like I said above, not a set and forget (but again I argue none of the stocks discussed on this board are set and forget)


“EVERYBODY” said “ANYBODY” can ship DVDs like Netflix and compete with them. Didn’t work out for anybody else, did it?

I’m not claiming SMCI is going to ever have NFLX-like results. My point is that, as investors, there is MUCH we don’t know about any business, including competitors. AND, to various degrees, its the same with every other business we invest in. (Sometimes repetition is good to drive a point home). Very few people have inside knowledge or true expertise.

That being said… let’s recognize that SMCI already has competitors. And those competitors are… competing. And golly-gee-whiz … SMCI is still growing very fast and making a lot of money. Will something transformative happen that changes all that?

Could be…

Q: But can anyone claim that some transformative change cannot happen with any other company? A = No.

That’s why any smart investor watches the results each quarter and stays reasonably aware of the news.

He is no fool who gives what he cannot keep to gain what he cannot lose.


Isn’t the cooling feature of SMCI’s servers supposed to be a standalone competitive advantage regardless of their partner relationship with NVDA? Or do I have that incorrect?