Wpr101's January 2025 portfolio review

Zen mind. Beginner Mind.

My portfolio at the end of January stands at,

AppLovin (APP) - 21.2%
Reddit (RDDT) - 18.5%
Astera Labs (ALAB) - 17.8%
Hims & Hers Health (HIMS) - 17.1%
Paymentus (PAY) - 11.4%
Credo (CRDO) - 5.2%
Natera (NTRA) - 3.3%
ACV Auctions (ACVA) - 1.3%
Perimeter Solutions (PRM) - 1.1%
Zeta Global (ZETA) - 1.1%
GeneDx (WGS) - 1.0%
Raspberry Pi (RPI.L) - 0.9%

Results
2024: +146%
2025: +9% YTD


The main changes to my portfolio this month include starting positions in a number of promising companies: Perimeter Solutions, Zeta Global, GeneDx, and Raspberry Pi. I added to my position in Astera Labs when it dropped significantly on January 27, and the sales came primarily from Reddit which reached all time highs. I would ideally like to get down to 10 or less companies after this earnings season depending on how the reports look for the 12 companies I own.

It is the first time I can recall the market getting turned upside down from an algorithm discovery like the DeepSeek model. Even though the R1 model was not released that Monday, the market woke up to the discovery. I learned recently the market drop stemmed from a single blog post or negative take on Nvidia where the article went viral. It seems remarkable to me that this huge market drop is literally from a single person’s opinion on how the future of AI will play out. Sometimes disruptions like this create massive opportunities for investors due to the panic.

It is times like these where it is nice to have a diversified portfolio that does not include only AI companies, even though I believe AI is the strongest secular tailwind I have ever seen. Somehow even with Astera and Credo dropping over 30% in a day and making up over 25% of my portfolio, I ended the month up 9%.


Another observation that is top of mind this month is that there are some newer SaaS companies which are largely going unrewarded or unnoticed by the market. Two of these companies are PAY and ZETA, and I will compare them to Cloudflare (NET), a company which I owned for a few years. I am using GAAP numbers here because it is a pain to look up each company’s adjusted numbers. Additionally, companies like Cloudflare gave out more egregious options packages in acquiring talent, so these are more apples to apples numbers. Comparing their last quarters,

PAY
revenue 232M, +52% yoy
EBITDA 14.4M
net income 14.4M

ZETA
revenue 268M, +42% yoy
EBITDA 4.8M
net income -17.4M

NET
revenue 430M, + 28% yoy
EBITDA -5.4M
Net income -15.3M

Just looking at these numbers alone, knowing nothing else about the companies I would expect them to have similar market caps. The reason being while NET has more revenue it is growing only 28%, and has negative EBITDA, while the others are growing faster and PAY has significantly better profitability.

Yet when we look at the actual market caps, NET is a 47.5B company, PAY is a 4B company, and ZETA is a 4.4B company. From my viewpoint this presents a massive opportunity in PAY and ZETA where they have less than 1/10th of the market cap of NET but should more likely be closer to an even valuation between these companies. I believe the reason NET has so much of a bigger valuation is because it is widely admired by the investment community as being a critical SaaS product that has a big analyst following. While PAY and ZETA are relatively unknown and not part of the in crowd with analysts.

I am using the earnings calendar planning to take advantage of this situation as my biggest holdings APP, ALAB and RDDT all report first. Depending on the results I will trim some of these as they rise after earnings, or in the case of a disappointment I could also sell a portion to reallocate to PAY and ZETA.


Reviewing my companies in order of allocation,

AppLovin (APP) - 21.2%
I am expecting APP to have a huge quarter as their web pilot is live and generating both revenue and net income already. Another factor with AppLovin is that they benefit from algorithm improvements which goes directly to their bottom line. They are one of the biggest purchasers of GPU clusters from Google, and any AI algorithm improvement strongly benefits them. I see this company as being a big potential winner of DeepSeek or similar discoveries.

Reddit (RDDT) - 18.5%
Reddit has been up 20% YTD and I have not seen much news. They have their Reddit Answers product in Beta which effectively seems like an AI that works on the Reddit corpus. I am expecting a beat for their quarter as a lot of their international translations should have gone live in Q4, providing more ad real estate. Additionally, the company has yet to step on the gas yet for their existing ad real estate as they have said that less than 10% of the comments section was targeted for ads. This is in stark contrast to Instagram or Facebook which are inundated with ads. I believe the CEO will use this lever smartly, ramping up the ads slowly, and this should ensure the company can beat any estimate they give as they can ramp up ads on will.

Astera Labs (ALAB) - 17.8%
Astera got absolutely hammered on January 27th and the stock is still down nearly a third from all time highs, even with a recovery going from $80 back to $100. I haven’t seen this company make a single misstep anywhere since their IPO and I used the sudden drop in price as an opportunity to add. One thought I had is that if models are able to run on smaller custom hardware this could benefit Astera. A critique of Astera from the analyst community has been that they do not have enough content in the NVL72 systems. I will be diving into their earnings soon but whatever impact DeepSeek has did not impact Q4.

Hims & Hers Health (HIMS) - 17.1%
HIMS has grown back into a top position by rising significantly towards the end of the month. I trimmed nearly a third of my position in regards to the Amazon news that they are getting into a similar business, yet the stock has nearly doubled since then. That was a mistake on my part to overreact to the news as the stock already traded down from $30 to $20 on the Amazon news, so a ton was negatively priced in. I recently learned that Roman a competitor is bigger than I thought as well with a similar non-public valuation to HIMS. I am cautiously optimistic about the company heading into earnings. They keep guiding to higher revenue percentage numbers every quarter, and I am looking for them to do the same this quarter to regain confidence in this company as a top position.

Paymentus (PAY) - 11.4%
PAY is down a bit since I started a position in them after their last report. However, the more I learn about this company, the more I like it and I wrote up some notes from their latest investor conferences. For this earnings I am looking to confirm the pull forward on large enterprise customers was not a one off fluke, but a more consistent trend of enterprises wanting to onboard the product faster. I have no real prediction about how they should do, but I am expecting them to outperform and be raising guides again.

Credo (CRDO) - 5.2%
Credo had an incredible guide from the last quarter with revenue projected to go up roughly 70% sequentially. While I am very bullish on the company’s prospects, my enthusiasm is tempered by the company being based in the Cayman Islands and having Asia as a major selling region I am less familiar with. In their S1 prospectus they were once fined for violating US export controls. Additionally, their product is lower margin than Astera’s. While it may sound like I am making a bear case, I am noting these yellow flags while balancing based on the absolutely enormous guidance they gave.

Natera (NTRA) - 3.3%
Natera is a company I would like to do a write up on soon, they do genetic testing and are marching towards profitability after reporting their strongest quarter ever. The company is already a ~23B company so my concern is towards valuation and that this may be already a fully recognized company by the market. I found a competitor in GeneDx (WGS) which is about a tenth of the size, and I have been debating which numbers and story I like more. This field of diagnostic genetic testing has a big tailwind and I like investing behind a strong secular trend.

ACV Auctions (ACVA) - 1.3%
I am skeptical the car market is recovering as the interest rates have not gone down as much as some predicted and the car market is highly dependent on financing for both new and used sales to consumers. I view this company as an innovator in a cyclical industry which depends heavily on macro conditions. While the company uses AI to analyze cars, I would not put them anywhere close to an AI play. Usually I try to avoid investing against cyclical trends. However, with ACV Auctions they are bringing a lot of new business to a previously stagnant market. My next steps will depend heavily on the numbers they report in this coming quarter as their guide was not that strong, but I got the impression they may be a sandbagger on guidance.

Perimeter Solutions (PRM) - 1.1%
Perimeter is a company I had looked into a few times as their last quarter revenue reached 102% yoy. They make solutions for wildfires both preventative and “retardant” solutions. My biggest critique of this company previously was that they depend on an unpredictable wildfire season which could be hit or miss. However, with these awful LA fires that came about in the middle of winter, I am making a prediction their business could benefit as governments will want to stock up and be prepared. I am impressed with the mission of this company where they say success in this business is measured in how many homes and lives they can save. Additionally, they seem to have the most environmentally friendly product that is available in this niche, making them the go to choice for governments around the world. Here are some additional thoughts about this company.

Zeta Global (ZETA) - 1.1%
Thanks to monkeydluffy who wrote up some thoughts on this company, I have been equally impressed by their last earnings call. The company’s revenue yoy growth rate goes from 20% → 24% → 33% → 42% surpassing the 40% threshold I typically look for on revenue growth. They are an AI powered marketing technology company that seems largely unknown outside of industry insiders. The company seems dramatically undervalued to be a SaaS company with 268M of revenue and ~4B market cap.

GeneDx (WGS) - 1.0%
I will be looking to create a separate thread on this company soon. GeneDx is a genetic and exome testing company that works with children and NICU patients. Their diagnostic tests have a huge benefit of finding rare conditions that could otherwise take years to find with multiple misdiagnoses. The current quarter of reported growth is 77M with 44% yoy growth, however they just pre reported numbers for their next quarter, and got 92M where they had guided for only 80M. This means the upcoming quarter has 59% yoy revenue growth and 20% qoq growth. The market barely reacted to this positive pre announcement and the stock even trended down after, leading me to believe this is a large opportunity here to get shares at a good value.

If you look at the chart of this company the price was massive in 2021 when they went public and absolutely cratered but then had a large runup. It is a bit more complicated story than I would typically like as they previously did a SPAC named Sema4, and then changed to GeneDx in January 2023. The company seems really on track now though as studies are proving to doctors the necessity of the tests to get good health outcomes in infants.

Raspberry Pi (RPI.L) - 0.9%
Raspberry Pi trades on the London Stock Exchange and is based out of the UK. You may have heard the name from the computer boards they make which were primarily for hobbyists and educators. However, as time passed industrial usage began to take hold and now accounts for about 70% of revenue with this number likely to increase over time. Surprisingly the company rallied big on the day of the DeepSeek news indicating to me that the market recognizes them as a beneficiary of AI which can run on small form factors like the Raspberry Pi chips. I gave a more detailed write up in this thread.


Companies I researched this quarter included,

Delcath Systems (DCTH)
I looked into this company after seeing the thread on the boards and this company is at the top of my watchlist now. Revenue is picking up nicely but still small at 11.2M, but the company only has a 500M market cap. They have one main product HEPZATO Kit which treats liver cancer. It works by isolating the liver, delivering chemo, and then filtering blood back into the patient. They have a strong expanding network of customers.

Bakkt Holdings (BKKT)
I saw that BKKT was number four on the IBD 50 list recently. The company does cryptocurrency custody and has a SaaS API for interacting with different markets. There is a Bakkt marketplace for trading, trust or custodial services. The strange thing is the company has a -5% gross margin, and I was very surprised that IBD’s proprietary screen allows in a company with a negative gross margin. This is one of the reasons I am critical of screens being used on their own, especially one that is proprietary and not hand crafted.

Comfort Systems USA (FIX)
FIX is a 20B company with 31% revenue growth on 1.53B of revenue, 170M in EBITDA, and 96M of net income. Those all sound like impressive numbers with an acceleration off smaller revenue growth previously. However, the company is an extremely stodgy company that does mechanical, HVAC, plumbing and piping services. They were founded in 1917 to show how old this company is. It is basically an amalgamation of different companies over the years moved into one larger holding company. While this seems like a well run business with accelerating financials, there seems to be little innovation here besides taking more market share and acquiring wisely.

Propel Holdings (PRL)
Propel is a lending platform providing credit products to underserved consumers. They are based out of Toronto and trade under several sub brands: MoneyKey, CreditFresh, and Fora Credit. The company has a market cap of 1.44B, with 117M of quarterly revenue and 41% yoy growth rate. Additionally, they are GAAP profitable with 22.8M of EBITDA and 10.5M of net income. The main drawbacks I see are they depend heavily on macro and underserved borrowers which is a nice way of saying junk credit. Additionally, the company just did an acquisition in the UK and plans to expand to Europe. While that could open up new TAM overseas, it is a risk adding in yet another brand under their umbrella.

Rush Street Interactive (RSI)
Rush Street is in the casino and sportsbook business operating in 15 states in the US. Internationally they are doing well in Ontario, Columbia, Mexico and Peru. They also have different online properties depending on the jurisdiction. Revenue is growing 37% at 232M and the company just reached profitability. The market cap is 3B, and this seems richly valued to me. This company has about the same P/S as SaaS offerings PAY and ZETA which get more consistent recurring revenue with a higher growth rate. I generally view online casinos as a net negative on society and different from skill games such as poker.

Rambus (RMBS)
Rambus is a semiconductor maker with a 83% gross margin. Revenue growth is an attractive 38% with 146M reported in the last quarter. EBITDA was 61M and net income was 49M in the last quarter which is incredibly high on the margins side. They were founded in 1990 and back in 2000 the company was $127 a share, it went down to $14 in 2020, and now sits around $56 a share. They make memory subsystems and get revenue from licenses of its tech. Their customers include data centers, automotive, mobile, IOT, and financial services. Market cap is 6B and they have 376M of cash with only 30M of debt. All of those metrics are impressive, but where I am very underwhelmed is on their guides and where they landed. The narrative says business is booming but this past quarter was guided for 144-150M and landed at 146M. Next quarter is guided for 154-160M which doesn’t seem to line up with the positivity on the call.


This month marks the first month of doing investing professionally and January 27th was actually my first market day where I am doing this full time now. I am feeling fortunate to have come out of that day unscathed and up significantly versus market averages so far on the year. I am optimistic this earnings season will go well, and I will be looking to cut any company which underperforms.

99 Likes

Thanks for the writeup, and good luck with the move to full-time.

11 Likes

I don’t think that’s the case. There was a “Leader” article in the Economist’s Jan 23 edition with the heading “Chinese AI is catching up, posing a dilemma for Donald Trump” with sub-heading “The success of DeepSeek and other Chinese modelmakers threaten America’s lead”. There was also a briefing (deepdive) article in the same edition titled “Why Chinese AI has stunned the world”. This is an excellent overview for those wanting to dive in btw. Those articles were written just before the weekend that predated the drop.

Also, after the R1 model dropped on 20 January the developer world (at least in my nick of the woods) took serious notice and started talking, downloading, experimenting with the model(s). Perhaps something like that takes a week to really sink in?

In addition to the developer buzz and outside of arguably the most trusted “mainstream” publication out there writing in depth about DeepSeek and other Chinese models and the implications of same, there was a small tsunami of articles, videos, opinions etc written about R1 after its release…

I, for one, wouldn’t notch half a trillion dollar decline in market cap to just one blog post. There was a chorus more.

Which leads me to this. There is a non-negligible chance that the chorus may just be right. And right does not mean that AI demand is going to collapse. Right in this context just means the AI capex needed and flowing to Nvidia to fulfill that demand could be lower than what the market expected prior to R1.

If nothing else, imo, these developments raise questions about the extent to which the level of announced and hyped AI GPU investment is going to be executed. Even if AI capex is half of what was being thrown around by the hyperscalers and others, that is still a gigantic number, bigger than just about anything that’s gone before. But if that number ends up being lower than what was expected prior to R1’s release, then the drop in NVDA’s share price could well be warranted.

I still own NVDA but I’m less sure that this is such a great buying opportunity.

-wsm

18 Likes

It could be a clickbait story from MarketWatch but here is what they said in the article, and linking to the blog post itself. The article is titled “The blogger who helped spark Nvidia’s $600 billion stock collapse and a panic in Silicon Valley”.

https://www.marketwatch.com/story/the-blogger-who-helped-spark-nvidias-600-billion-stock-collapse-and-a-panic-in-silicon-valley-52aba340

It seems the main conclusion from the blog post is this,

I think you are right it is a confluence of factors that alerted the market to the changing dynamics of the AI algorithms. Although it does seem like this one guy’s blog post going viral may have tripped off some institutional selling. As I’ve ramped up on learning about DeepSeek a few people have pointed out the big break through really came with DeepSeek V3 released over a month ago, and a few people back then pointed out the massive breakthrough.

I believe trading dynamics also played a large role as many momentum investors and institutional investors use stop losses generally around the 10% range. When a stock crosses through losing 10% on a given day often it turns into a falling knife scenario which explains how some of these semiconductors lost 30% in a day.


META has already come out and raised their CapEx prediction for 2025 which was originally slated for 50B, from their conference call on January 29,

Turning now to the CapEx outlook. We expect our full year 2025 capital expenses will be in the range of $60 billion to $65 billion. We expect CapEx growth in 2025 will be driven by increased investment to support both our generative AI efforts and our core business. The majority of our CapEx in 2025 will continue to be directed towards our core business.

Could it be a waste of money for them? Possibly, but it still seems like the race is on. It is debatable though if they will get a decent ROI or not from that spend though.


One thing I like about Astera specifically is they are agnostic about what the customer is building and can adapt to the trend. Effectively a hyper scaler like META tells them what they need and they look to build it to spec. They can help diagnose and troubleshoot for custom hardware solutions that hyperscalers are building out.

21 Likes

I think the reason networking companies like Astera, Arista, and Broadcom dropped more than Nvidia is that if large clusters aren’t needed, then fewer interconnects between servers are needed. Add in for Astera that Nvidia has redesigned its Blackwell servers to have shorter data paths, and so fewer retimes are needed.

Similarly, if large clusters aren’t needed, then the push for more power also isn’t needed (POWL, VRT, etc.).

So, Nvidia would still sell chips, but the large data center specific infrastructure around them just isn’t needed.

It is true that Astera’s retimers and PCI networking could be needed for the non-Nvidia, in-house developed ASIC chip development being done by Amazon, Google, etc. But again, if large clusters aren’t needed for training, then it doesn’t matter if the clusters are Nvidia Blackwell or Amazon Trainum.

This is all if you believe that is the impact of DeepSeek. We just saw Meta up their AI capex from $50B to $65B for this years. So, maybe not.

35 Likes