Wpr101's portfolio April 2024

Zen mind. Beginner mind.

My portfolio as of today stands at,

Super Micro (SMCI) - 19%
Nvidia (NVDA) - 16.2%
Elf (ELF) - 16.1%
TransMedics (TMDX) - 15.5%
Celsius (CELH) - 14.2%
Axon (AXON) - 7%
AppLovin (APP) - 6.4%
Hims & Hers (HIMS) - 5.5%

This month I sold four stocks: IOT, BEEM, NET, MNDY. I added one new company AppLovin (APP). I added to my positions in NVDA, TMDX, and ELF, and lowered allocations in HIMS, and AXON.


This month I was working with a mental game coach on investing and trading. The two most prevalent errors that I am making are copy trading and FOMO based trading.

Copy trading could be as simple as doing slightly less research on a company than you normally would because investors you respect have recommended the company. This tactic effectively undermines the entire investing approach of being in your top conviction positions, you are in someone’s else top conviction position. I strongly agree with Saul and Bear have said before about not copying their trades - the only thing you are doing is harming yourself by copy trading.

To counterattack the tendency to copy trade, for companies in which I learn about the ticker directly from someone else recommending them, I am doing more even more research on them than I typically would. In this way I won’t have a dependency on the recommender to have to follow their updates on the company.

With FOMOing into trades, this happens to me when I first learn about a company and show recency bias of preferring the new company to existing companies I know more about. With how fast some of the AI company stocks are rising right now, it’s easier than ever to have this tendency. To counterattack this tendency, if I am really interested in a company but haven’t done enough research yet I’ll take a small starter position.


Currently I have two companies on my watchlist which are Tidewater (TDW) and Aspen Aerogels (ASPN). The ramp up time to learn about these companies is quite large for me as these are industries I am completely unfamiliar with. I may start a position in small positions in these two companies as I learn more.

Reviewing my companies in order of allocations, my sells, and then companies I researched.

Super Micro - SMCI 19%
This is a very high conviction position for me. I believe this company has a significant competitive advantage in building AI servers. Facebook (META) is their biggest customer and going to be investing over 100B in AI over the coming years. This company is an engineering powerhouse. I wrote up some thoughts on their last earnings here

I’m holding at these current levels, because they were supply blocked this quarter, and I’d like to see them a deliver an outstanding quarter before wanting to add any more.

Nvidia (NVDA) - 16.2%
I am very confident in Nvidia and their future prospects. Watching Jensen’s 1 hour 45 minute GTC conference was like seeing a movie. I’m more impressed than ever with the visionary leadership this company has. They have evolved way beyond their gaming company origins - so much so that the GTC speech never mentioned gaming once. Seems like Jensen’s interests are towards optimizing industrial operations.

As Jensen points out, this AI industrial revolution is much different than the CPU revolution. In just eight years, Nvidia has increased computation by 1,000x. By comparison Moore’s law was roughly 100x every 10 years.

Yet as Jensen points out, it’s 1000x already by year eight, and “there’s two more years to go”! Jensen said, “the rate at which we are advancing computing is insane”

Here you can see the in graph in 2016, Pascal the fastest chip was 19 TFLOPS on 16nm. Now the fastest chip is Blackwell at 20,000 TFLOPS and 4nm. This is the 1,000 times computation increase in just eight years.

Elf (ELF) - 16.1%
I like this company a lot, and how their revenue is accelerating along with bottom line growth. I’ve added to my position recently as I don’t see that worry of Ulta sales being a big issue. Looking in that, it seems like Ulta is a dying business, but I’m not sure that significantly impacts Elf. At $160 this price seems quite reasonable to me.

TransMedics (TMDX) - 15.5%
I was writing that I thought at $70 TransMedics was a solid buy. I added a little bit around 80, and then added again after they reported earnings. I liked this earnings a whole lot and wrote up some thoughts here.

Celsius (CELH) - 14.2%
I haven’t done a lot of research on them recently but still fairly high conviction here, I might like to add a bit at these $75 prices. I’d expect them to do well on earnings.

Axon (AXON) - 7.0%
Axon is right on the edge of where I’d be interested from a revenue growth perspective of being at 30%. I still have some questions where why their product is not taking off internationally. The financials look really good. My confidence in the leadership team is not that high though.

AppLovin (APP) - 6.4%
There was a great thread started on AppLovin over here. I’ve added my thoughts there and I am really interested in this company. I’d like to live through at least one earnings and see how they do before I re-evaluate.

Hims & Hers (HIMS) - 5.5%
I decided I wanted a slightly smaller position going into the earnings to see if their net income can continue to accelerate. This is still a small company, and has more risk than others. Wrote up a lot of thoughts over here


Companies I sold: NET, IOT, BEEM, MNDY

I’ve become more skeptical of the SaaS business model in this ~5% interest rate environment. It seems like most of the companies now end up trending to 25% growth over time.

Cloudflare - NET
This company has become a disappointment for me. I’ve been invested since their 2019 IPO and followed closely. This company seemed like a bargain at a 5B market cap for a company which was said to “run the internet”. In fact, I remember before this company went public on Hacker News it was commonplace for people to say “If Cloudflare is down, the internet is down”.

Unfortunately, the way I see it now, this business is run like a charity, where doing good is more important than shareholder return. This is really an incredible product suite they offer, but if they are unwilling to monetize it anytime soon, I’d rather be out. I could become interested again if the price gets really low, or they somehow figure out a way to get their sales organization together.

Samsara - IOT
My biggest concern here is that 2025 revenue is projected at 30% and they project employee growth to be 30% that year. Maybe it seems like a weird stat to harp on, but I figure if they grow their headcount by 30%, revenue should be growing much faster than that. Plus I have concerns their product may not have as big as a competitive advantage as I thought - they seemed to misplay the situation with that lawsuit.

Monday - MNDY
This company has very promising financials, but it seems like they have on and off again quarters. They don’t seem to be in a rhythm on acceleration of revenue. A couple quarters ago I was really impressed with what they said on the call. Unfortunately I have macro concerns about this company being based in Israel even though I know most of their customer base is US enterprises.

Beam Global - BEEM
I have so many red flags written down for this business it’s not even worth going into them all. First earnings was delayed by over 2 months and they should already be reporting again soon - the reason: they took on too many new initiatives. Next their sequential growth is not that impressive because they are including the Serbian company.

Biggest red flag of all it’s the CEO’s comments about these warrants they had advertised on their investor relations page. These warrants had a strike price of $6.40 and the stock price closed that day at $6.30. The CEO said the warrants were in the money, but this was just an outright lie. Also how does he consider the warrant being at the money to be worth anything? There’s only two possabilities - he either doesn’t understand what a warrant is, or there’s something worse going on.

Energy Services of America Corporation - ESOA
I am fascinated by this company and held a tiny position in them before selling. They have a 14x PE and revenue is growing 50% year over year. Current revenue in the last quarter was 90M dollars, but the market cap of the company is 114M.

They do energy services along the East Coast. I’m not sure there’s much innovation here but this company seems tremendously undervalued - or maybe I’m missing something completely here. They don’t even do earnings calls and my understanding this is run a bit differently than a typical wall street stock.

DocGo (DCGO)
This one is a very complicated story. Revenue is almost 200M and grew 80% yoy, the P/E is 70 and this is a profitable company. Yet the market cap is only 368M.

The old CEO’s name is Anthony Capone or Tony Capone - no joke. He resigned in scandal, but what was the scandal he was guilty of? They said he lied about which college he went to on the homepage of the website. He went to some lesser accredited university but it blew up into a scandal.

The company also had a number of problems with the New York city about a contract to handle migrants. This company provides healthcare to people in need to stay out of the ER. They run ambulances, full staff with doctors and nurse and mobile ICU units basically. They can potentially step in for disasters and help out - the mission seems quite nice. However, the City of New York claims they were badly in breech of a no bid contract. Failing many responsibilities they had. My take on the situation is the migrant situation in New York spiraled out of control - beyond what anything DocGo anticipated.

Long story short, I invested briefly in this company but then sold my shares short after that after going down a rabbit hole here.


Companies I researched,

ASML Holdings
Was checking in on what this company does and if it has relevance to the semis I follow like SMCI and NVDA. Not enough growth here, but sounds like a decent enough company. Put some thoughts on this company in a thread dedicated to the industry.

Arm (ARM)
This is a really interesting British company in the semis market. I’d like to learn more, sounds like their licensing of designs is an interesting business model.

Universal Technical Institute (UTI)
They provide secondary education for students seeking careers in professions that do not require a degree. Growing fast, nice financials. Some criticism of the business online for charging too high of tuition.

Tradeweb Markets (TW)
Financial players for trading bonds and other assets. Hugely benefitted from higher rates, the market cap on this company is very big though and its not growing fast enough to be interested more.

Duolingo (DUO)
I finally got around to reviewing DUO. It seems like a decently run software company and they may be able to get into some other learning markets. I just cannot wrap my head around why this app is so popular - and the whole business is based around the app. I tried using the app for a few weeks and it was good, but it wasn’t something incredible I cannot stop using.

StoneCo - (STNE)
Well run Brazilian financial and looks like nice growth. They don’t offer guidance. I’m not thrilled about investing in Brazil.

MercadoLibre - (MELI)
Seems like a well run company with a good leadership team. I think this company will likely do well, although I think they will face more competition. The board knows much more about this company than I do, but some things leave me with questions about this company. Specifically, I’d heard they were completely dominant throughout Latin America but it seems 95%+ of the revenue comes from three countries. The ramp up time on this business to learn the ins and out seems enormous, in addition to following individual country risks, this one is not for me.

Pinduoduo (PDD)
I’ve been hearing about this company Temu over and over again in different podcasts and seen their ads targeted at me. I’ve been wondering what the heck these ads are for these knockoffs items and prices that are so low they don’t make any sense.

Turns out Pinduoduo owns Temu and revenue is growing 123% while EPS grew 143% year over year on their last quarter, the P/E is 23 which seems reasonable. They are a Chinese company though, so they are going to trade at a discount because the risk of fraud is so high. I was also comparing PDD to Kaspi (KSPI) and MELI.

I find Kaspi to be the most appealing of these three. Yes, investing in Kazakhstan is extremely risky, but this company has a P/E of 10, pays a 6% dividend and has a 45% net income margin.

Lantheus Holdings (LNTH)
They have tools for diagnosing cancer and heart disease. The have very strong bottom line growth and revenue growing the in the mid 30s. This story requires a big ramp up - complicated tools to understand the conditions they treat. I saw they had a bump after earnings, so may look again.

Li Auto (LI)
Growing revenue 136%. Chinese car market, gross profit growth and EPS growth are amazing. If this was an American company with these numbers I’d be all-in. Again too concerned with the Chinese markets which are rife with fraud.

Navitas Semiconductor (NVAS)
Fast growing semi, they make GAN circuits and silicon carbide tech. Ultrafast charging for mobile, electric vehicles, and industrial applications. Large customer base of 10 of the top 10 OEMs. Pretty small company. I had some concerns they are guiding to be down sequentially in revenue.


Wrapping up three of my companies reported earnings so far and I’m pleased with the results of SMCI and TMDX. Both companies say they can accelerate revenue sequentially for the foreseeable future!

I’m happy to sell Cloudflare and re-allocate to my other highest conviction positions.

Next week I’ll be looking at a lot of earnings reports again - HIMS, CELH, AXON and APP

61 Likes

That sure is a concern for me, and has kept mealy from reinvesting in it.
Saul

16 Likes

Related to the comments above about NET - is it possible that their pricing strategy is designed to keep out competition? This tempers revenue growth in the short term, but sets them up for long term success. Kind of the Amazon model. IDK - just brainstorming.

I still own it, and it’s my second largest individual stock investment at ~ 2.4%.

The price drop is not fun, but it just got over-extended on the upside and its been correcting. I thought the results from the ER were great. Brad Freeman wrote a good summary if anyone is interested. His conclusion is that it’s an elite company, but the stock is too expensive for him. Brian Stoffel had a good video summary on X also for the recent earnings.

13 Likes

For sure, this seems accurate. Their basic website DDoS protection service is free for any website. They even have a free landing page, Free Plan Overview

The company used to report on the number of free customers they had as well which was massive. That was part of my original investing thesis, that they have this huge funnel of startups and other websites to sell into. The idea being as these startups gain traction, they will ramp up traffic and upgrade to paid plans.

I don’t see how many company could compete in this area with Cloudflare they are the defacto solution for DDoS and basic DNS protection.

I still see Cloudflare as a company with leading edge products but I don’t think they are priced appropriately. About a year ago they started posting huge year over year, and quarter over quarter numbers on their Workers program, and said how many use cases there are for AI.

I assumed Cloudflare would benefit massively from AI as they are saying they have a top platform for building out AI applications. However even with all the growth of Workers it’s still not really adding to revenue. The program is far out of Beta now, so it’s not like it should be getting run at cost now. Another concern is if they can raise prices later if their customers get used to low prices.

They are gaining market share in a lot of areas by out competing on price, I just don’t see it as a great business model currently in this ~5% interest rate environment.

Lastly, I think there will be time to get back into this company if they do hit the accelerator on monetizing. Maybe this will turn out be like Amazon where they built out this next generation AI infrastructure that becomes indispensable to the internet and AI. I think there will be a lot of time to potentially get back in if this is the case.

I could be interested again in Cloudflare if the AI portion of Workers really starts to ramp and they break out that revenue to report on its consistent growth. Another factor where I could invest again is if they out innovate ZScaler, offer a vastly superior product, and start taking more share in Zero Trust. I’d want to see that they sell Zero Trust for more than just at cost though.

13 Likes

Both very fair comments. The flip side is NET has been asking for (and getting) the benefit of the doubt on both these ideas for multiple years now, but we have yet to see even one breakthrough product being significantly monetized.

Cloudflare is no longer a 50%+ grower at a $750M run rate. It is now coming up on being a sub-30% grower at a $1.6B run rate. Being honest, we have several companies at the same or larger scale showing both better top line growth and bottom line leverage.

In the concentrated portfolio many here run, that matters. At some point, the opportunity cost of owning a company becomes greater than the opportunity. Cloudflare is clearly at the point where investors are choosing sides…and that’s what makes a market.

28 Likes

Wpr101, First thanks so much for this outstanding report, including links to your detailed analysis. Your ownership of stocks overlaps a lot with mine - it’s nice to see validation and it’s also nice to see where we are different so I can learn from new thinking. And I certainly could get better about copy and FOMO purchasing, it’s nice to see it called out as something to not just know about but address.

I want to comment on the above quote though, because it’s the second time I"ve seen the “why doesn’t their revenue come from more countries” question about MELI and since I think it’s an outstanding company, I’d like to suggest a reason.

Attached below are the first 20 countries in Latin America,/Carribean descending population order. (Sorry about the ugliness and excessive information - could not figure out how to attach a selective picture/screenshot, if someone can tell me which little icon will let me do that I would appreciate it)

As you can see, Brazil, Mexico, Columbia and Argentina are massive markets, followed by Peru and Venezuela (which is in economic hell right now) and the population drop off is super steep after that.

Chile, at number 7, is only 1/10 the size of Brazil!

Now, if you cross that with relative wealth, you can see why the majority of sales come from the three countries. Even though MELI operates in 18 countries, population and per capita income per country easily explains almost all their concentration.

Top per capita income: Uruguay (Less than 2% size of Brazil ), Chile (10% size of Brazil), Guyana, (population 814k), Argentina, Mexico, Brazil.

If I were MELI, I’d really work on those big markets before I worried about picking up more from the little players.

There may be other reasons you have to stay out of MELI but I don’t think this should be one of them.

1 Brazil 216,422,446 0.52 % 1,108,948 26 8,358,140 6,000 1.6 34 88 % 2.69 %
2 Mexico 128,455,567 0.75 % 951,442 66 1,943,950 -50,239 1.8 30 88 % 1.60 %
3 Colombia 52,085,168 0.41 % 211,144 47 1,109,500 -175,051 1.7 32 81 % 0.65 %
4 Argentina 45,773,884 0.58 % 263,566 17 2,736,690 3,718 1.9 32 94 % 0.57 %
5 Peru 34,352,719 0.89 % 303,131 27 1,280,000 -61,442 2.1 29 79 % 0.43 %
6 Venezuela 28,838,499 1.90 % 536,803 33 882,050 321,106 2.2 28 N.A. 0.36 %
7 Chile 19,629,590 0.13 % 25,857 26 743,532 -71,205 1.5 36 85 % 0.24 %
8 Ecuador 18,190,484 1.05 % 189,484 73 248,360 -21,525 2.0 28 64 % 0.23 %
9 Guatemala 18,092,026 1.39 % 248,118 169 107,160 -9,110 2.3 23 55 % 0.22 %
10 Bolivia 12,388,571 1.35 % 164,461 11 1,083,300 -3,000 2.5 24 69 % 0.15 %
11 Cuba 11,194,449 -0.16 % -17,742 105 106,440 -6,000 1.5 41 80 % 0.14 %
12 Haiti 11,724,763 1.21 % 139,767 425 27,560 -31,811 2.7 23 60 % 0.15 %
13 Dominican Republic 11,332,972 0.93 % 104,151 235 48,320 -29,099 2.2 28 85 % 0.14 %
14 Honduras 10,593,798 1.54 % 160,938 95 111,890 -5,034 2.3 24 58 % 0.13 %
15 Nicaragua 7,046,310 1.41 % 97,918 59 120,340 -8,000 2.3 25 56 % 0.09 %
16 Paraguay 6,861,524 1.19 % 80,780 17 397,300 -12,499 2.4 26 67 % 0.09 %
17 El Salvador 6,364,943 0.45 % 28,551 307 20,720 -23,249 1.8 27 78 % 0.08 %
18 Costa Rica 5,212,173 0.60 % 31,344 102 51,060 3,750 1.5 34 82 % 0.06 %
19 Panama 4,468,087 1.35 % 59,506 60 74,340 7,262 2.3 29 70 % 0.06 %
20 Uruguay 3,423,108
24 Likes

I’m just curious because I noticed that you have ELF, both use low-price strategy. Why do you worry about NET but no worry ELF?

I’d say there’s two noticeable differences differences, the first being that Cloudflare gives away free products while Elf does not. The other is that Elf is profitable on the bottom line from a GAAP perspective while Cloudflare is not.

To give a couple examples, the CEO of Cloudflare used to talk a lot about the product “1.1.1.1” on past conference calls. This is a free VPN service that can be used at no cost, and I’ve had it on my phone for years. There is an upgrade to an even faster version of the VPN which is $4.99 a month. However, I doubt the free version plus the revenue of subscribers covers the cost to even run this application.

That is in addition to have built the 1.1.1.1 application, have engineers do maintenance and troubleshooting, and then providing customer support to paying members.

Also, this was the first conference call I listened to the audio in a few reports, and I could hear how excited Prince was by getting “tremendous goodwill” from the community for Project Athena, which is providing free services to the White House and many states. I’m not clear at all why the White House and state governments are not paying anything for this service? Surely the US government has budget for cyber security.

Elf on the other hand is selling a low price product that generates GAAP EPS. If you check out “prestige” brands compared to Elf they both have gross margins in the 70% range. I don’t think there is any product Elf produces which is sold at cost, let alone given away for free.

32 Likes

I think you are misunderstanding the purpose of these free/goodwill projects that Cloudflare services for vulnerable/non-profit groups. The Athenian Project is free security service provided for State & local government election websites (Athenian Project | Cloudflare). This is likely a very small amount of their network traffic and miniscule cost to them. Cloudflare is knocking up on 80% gross margins, where I assume these costs would fall into. Cloudflare does not provide free security/software services to federal or state governments outside of these election websites. This is Prince on the Q1 call on how this goodwill has benefited them in driving business and significant revenue from the government:

Matthew Prince

Yeah. I think federal is exciting for us. Last year, we achieved FedRAMP certification. That’s opened a lot of FedRAMP-- excuse me, of federal deals to us. I think we’ve always had a great relationship with a lot of the members of the federal government. I think the thing which is hard to appreciate that, as I talk to our federal customers, is how much they appreciate that when they ask us to help with something, we step-up and help. So, this is – this year in the United States, obviously an election year, we’ve been running the Athenian project since 2016, which provides our services at no-cost to anyone who is helping administer an election anywhere in the United States. We’ve done similar things around the rest of the world. We’ve worked with the White House on a number of initiatives, including Project Safe schools in order to protect the most vulnerable schools and communities across the country.

And that is all the foundation, which I think has built an incredible amount of trust and an incredible amount of goodwill in the federal space. And so today, as the world is getting scarier and scarier, and we’re seeing more and more cyberattacks, the federal – our federal business, I think is being driven by the fact that we have great products, the fact that we have the certifications that are required to serve the federal government, and the fact that we’ve built an incredible relationship as being not just a trusted vendor, but a true partner where we have helped make sure that when there’s any public institution that is in true need anywhere across the country, we step-up and help. And I think that’s now turning into more and more business that is also generating significant revenue for us.

I disagree with your thoughts that these goodwill projects are a bad thing in the long run for the company for these reasons.

Bnh

22 Likes

What cloudflare does is basically product-led growth. I love seeing companies doing that, since it’s a very smart way to get more customers in - as long as the product is great.

Monday is another business getting customers in by using that strategy. I wrote about it here: Monday.com reports earnings - #17 by mooo

12 Likes

I think you are misunderstanding the purpose of these free/goodwill projects that Cloudflare services for vulnerable/non-profit groups. The Athenian Project is free security service provided for State & local government election websites

Quoting directly from the Athenian Project landing page linked,

Free security and performance for state and local election websites

Athenian Project gives our Enterprise-level protection and performance for free.

I am questioning why “enterprise level” software is being given away for free. The State of California falls into the category of “vulnerable/non-profit groups”?


Is there any other SaaS company that is giving away their Enterprise level product for any reason?

Only example I can think of are the large cloud providers like Oracle who gave Zoom two years free of using their servers. However, Oracle is the same company that is known in the industry for turning the screws on it customers sending teams of auditors and lawyers to squeeze the maximum out of their customer’s billing department once they have vendor lock-in.

I’m seeing Cloudflare optimize for goodwill over operating profit, whether this is their free tier DNS/DDoS protection, 1.1.1.1 product, and Athenian project. Even for Zero Trust they say they bundle Zero Trust to customers for basically no additional cost!

They’ve said Zero Trust is a product that has “well above” 90% margins, but what good is that to investors if the product is sold at or below cost?

8 Likes

You partially answered your own question with the last sentence. Because of the way the network is architected and because Cloudflare owns the infrastructure and can manage the traffic it is done at very low (essentially 0) cost to them. You ask why other SaaS companies don’t do this? They don’t own their own infrastructure, they rent it from the hyperscalers. It would be a lot more expensive for them to provide something like this. This is a major advantage Cloudflare has over other SaaS businesses.

Why does Cloudflare give these products away for free besides being almost no cost for them? Prince explained that it has contributed to building trust and goodwill in the federal space and is turning into more and more business and generating significant revenue for them.

Outside of the federal space the free tier of their products allow for rapid product development in that they can release new products and use the free base as testing/QA on them. It also make it very hard for competitors to compete with. There is a very strong moat around the business and this is part of the reason the valuation always has been so high despite slightly more tepid growth than other similar hypergrowth Co’s.

This is an answer directly from the CFO at the Morgan Stanley Tech conference in March this year when asked about why they have a competitive advantage in SASE vs. others, but it also explains why this can be done at almost no cost to them. It is pretty clear to me that the benefits of this “free” software outweigh the costs.

And every product we have and every service we offer runs on every server and every location and that means that the complete surface of the network capacity wise and infrastructure-wise, becomes our decrease of freedom, how we manage traffic, how we manage cost. And this the key reason why our margin structure is so superior and why you have such an elasticity in our business model.

So for example, when – during COVID when most of our revenue is subscription review, it’s also quasi fixed, there’s very little component of variable. When during COVID, we all started to work from home, traffic spiked on our network literally within a couple of weeks by 60%. Folks expected our margins to tank, they didn’t flinch, they actually improved. And this speaks volumes about the efficiency of the architecture but also the elasticity we have to absorb gigantic moves in data.

Now this network is built on the traffic we deliver or handle with our first wave of products. So it’s a CDN network, but not a lot of CDN revenue, it’s the firewalls, the DDoS mitigation, the routing, the load balancing that happens. In this business model, you pay not – or we don’t pay for the amount of data we move, we literally pay for the size of the pipes we have installed, right? And the first wave of product is literally traffic moving out to the eyeballs.

So when we now design a slate in the portfolio of Zero Trust products, they are literally moving traffic in the reverse direction. It’s all about moving traffic back. So all that traffic that we collect literally comes for free. So our Zero Trust products are margin-wise far of 90% there. Matthew, I think, on the last earnings call [indiscernible] he said we could consolidate all the Zero Trust providers out there, put all that network on our – all the traffic on our network and not need to invest one additional dollar of CapEx. So it gives you really a good idea of that – the capacity of the network.

22 Likes

First, thanks for the great thoughts on all your companies and more. This is a very nice April review and I’m glad it has already sparked some discussion.

As for SMCI, I guess I’m a little surprised your conviction is “very high” and that you’d even consider adding more to a very large 19% position if they deliver an outstanding quarter next time around. I guess I’m thinking that it’s really hard to predict what revenue and profits will be like going forward. If they keep the same 9% net margin they had in 2023, profits in 2024 will be ~$1.35b and they will need a fwd PE of about 35 to even justify the current $46b mkt cap. Of course, if they can grow revenue this fast in 2025, that is cheap. But is anyone in the world expecting them to put up ~$30b next year after doing ~$15b this year? What do you expect? I think that’s the difficult thing with this company. Revenue has boomed the last few years and 2024 appears like it will be their best by far. But if you look at 2018-2021, there wasn’t any revenue growth. So it just seems unpredictable to me. What gives you such strong conviction?

Bear

13 Likes

Are we talking Fiscal or calendar year? Because if you are talking Fiscal then the Profits should be 2.04b according to the analysts who I think can’t keep up with how fast this company is growing.

Andy

6 Likes

Thanks, Andy, that’s my bad – I was conflating the two. Stupid June ending fiscal year!

So this June for F2024 analysts expect about 15b revenue and $23 or $24 EPS.

For F2025 (ending June 2025), they expect ~23b revenue and $33 or $34 EPS.

For F2026 they expect maybe $27b revenue and $42ish EPS…but these numbers are based on fewer analysts wanting to go out that far and are pretty close to wild guesses.

If you’re going on these, I think the company is arguably pretty fairly valued. The slowdown to just ~17% revenue growth would be severe and the PE would probably contract.

It’s really hard to know for a company like this how much revenue will repeat steadily or even grow. I think to look at all the possibilities we have to consider that they could grow much faster (maybe 40%) in F2026…or they could grow none or even see revenue fall.

That’s why I think we’re seeing such volatility. If you plug in 40% maybe it’s worth $1000+ where it was at recently. If you plug in negative revenue growth it’s maybe not worth $500.

My point isn’t that I know which of these is most likely. My point is that I don’t see how this can be a high confidence position.

Bear

10 Likes

All good point Paul but they are bringing on line 2 more Plants in silicon valley and one plant in Malaysia all this calendar year(Yea those off quarters suck, makes it harder to discuss)

But with that extra capacity how do you figure their growth? It makes it hard because we don’t know how soon they will ramp. Right now they are doing about 2000 racks, liquid cooled which is what AI is going to, per month. They have 4 billion in inventory and they are projecting 5.5 billion in Revenue growth. I don’t see how they can’t blow that number away.

Andy

5 Likes

I would like to add that Charles Liang commented in his prepared remarks during the recent 3Q24 CC, “We like to support strong short- and long-term growth with minimal equity dilution. Overall, I remain optimistic that AI growth will continue for many quarters, if not many years to come.”

And in addition to the capacity expansion they are also scaling up the facility in Taiwan.

Yeah, I’m also long SMCI with about a 20% position.

11 Likes

First, thanks for the great thoughts on all your companies and more. This is a very nice April review and I’m glad it has already sparked some discussion.

Thanks! I’m glad to see a lot of discussion as well, both bull and bear cases on a lot of the companies the board is following.

Probably should have caveated the statement about adding that if they report well but the price doesn’t correspondingly rise to what I think it’s worth. 20-25% is really the max of what I would want it to get to, but would be heavily dependent on where it trades and what they reported.

What do you expect? I think that’s the difficult thing with this company. Revenue has boomed the last few years and 2024 appears like it will be their best by far. But if you look at 2018-2021, there wasn’t any revenue growth. So it just seems unpredictable to me. What gives you such strong conviction?

The generative AI moment hadn’t happened in the 2018-2021 timeframe so there was no such product as an AI server. They are saying themselves that they will be able to grow sequentially for the foreseeable future, including seasonality. CEO Liang says,

But now with AI, we’ve been growing so strong. So we basically are able to grow sequentially. So although March and September be a little weak, but basically, because of strong AI growth and our market share growing, so the sequential growth will become the normal.

And next quarter their guide is massive, they raised it by 400M to 5.1 - to 5.5B in revenue, and 7.62 - 8.42 of EPS. Keeping in mind their most recent quarter was their biggest ever at only 3.85B revenue, and 6.65 of EPS.

The biggest variability comes from supply issues which showed up this quarter. I like that they booked the huge incoming supply they got on the last week of this past quarter. I think it would have been very tempting from the CFO’s perspective to account for that inventory landing next quarter when the product will actually sell because it torpedo’ed the cash flow and margin numbers for this quarter. The company is very forward looking and ambitious.

Indeed, our brand is very ambitious. Let me use that word. We have a very ambitious brand. So we try to continue to grow very strong, kind of 3x to 5x faster than the industry’s average.

This is finally a company I’ve seen where the narrative really matches up with the numbers they are putting up. They are raising capital to expand, but it’s going for good purpose to drive both top and bottom line.

It’s really hard to know for a company like this how much revenue will repeat steadily or even grow. I think to look at all the possibilities we have to consider that they could grow much faster (maybe 40%) in F2026…or they could grow none or even see revenue fall.

Their biggest customer is Meta whose stock dropped because of how much CapEx they are putting into AI. They said they are going to spend 100B on AI over the coming years, and almost all of Meta’s last conference call was focused on AI.

Is raises the question why Meta with it’s huge budget is buying from SuperMicro if this was an easy to replicate product they could just build in house for cheaper. My thesis is it’s not worth the time or effort for Meta to build these servers so they’d rather purchase from Super Micro. If Meta is not going to build their own at the scale they run at, I cannot see how it would be worth the effort of another company to try and build AI servers themselves.

This was a really informative interview with Zuckerberg about where AI stands. They are open sourcing their Llama 3 model to compete with the likes of OpenAI and Perplexity, but yet they are going to Super Micro for the actual compute and data center build out.

20 Likes

What scares me about this thesis is it could have been the same for Aehr – just insert their name instead of SMCI and insert “On Semi” instead of Meta.

Anyway, just food for thought, nothing more. I don’t have a position and I find it interesting that many others here (e.g. Saul) don’t either. I’d be curious what they think too. We all want to find companies that will grow fast for a long time…some of us aren’t convinced on this one and I think it’s worth discussing.

Bear

15 Likes

I’m not really seeing the similarity to Aehr here. Onsemi accounted for 80% of Aehr’s revenue while Meta is 21% of Super Micro’s revenue. Super Micro has stronger demand than supply, so if Meta left as a customer it wouldn’t have that much impact revenue wise because those same servers would get sold to other customers demanding them. If OnSemi left Aehr the company is practically finished as they don’t have other customers to fill the demand.

The thesis I am mentioning is not about Meta needs to be their customer to be successful. It’s an observation that Meta, at their scale, is not building these servers themselves. To me that says a lot, because Meta is competing to build the AI models themselves with Llama, but they don’t see room to compete or out innovate Super Micro.

18 Likes