Argosy Investors’ commentary for the fourth quarter ended December 31, 2025.
Read more hedge fund letters here
Dear Investors,
I want to first apoloize for the delayed publication of this letter. The investment world, specifically related to AI releases from Anthropic, Gemini, and OpenAI, has been digesting seemingly daily changes at a rapid pace, and causing stock price declines for quality businesses at a frequency and magnitude that has not happened in a long time.
Looking back on 2025, the market was defined by developments in artificial intelligence (AI). Perceived AI winners gained significantly, even as some of the luster on AI came off during the second half of the year. Perceived AI losers were and continue to be discarded with impunity and without much discretion; there are likely some babies thrown out with the bathwater there. Companies perceived as minimally affected by AI positively or negatively largely treaded water as well. Many of my investment choices fall in this 3rd category of minimal affect from AI, and my largest winners this year reflect that.
With all that said, 2026 is off to a chaotic start. There seemingly are no longer any stocks without any impact from AI. Recent AI releases continue to show true use cases for white collar work as varied as finance, law, software development, and insurance/real estate brokerage. Several software companies in particular have experienced significant declines, as the market works to re-assess the impacts of AI. Right now, virtually everything software is being discarded until either enough time passes or other catalysts arise that disprove the current belief that essentially all software will be impacted. This period reminds me of several years ago when Amazon’s retail dominance seemingly knew no limits. Seemingy all retailers from grocery to hardware to electronics were going to be dominated by Amazon imminently. While Amazon continues to make progress and push into more verticals, those fears are now 10+ years old.
Returning to software, there appears to be at least a few vectors that could negatively affect valuations: 1) fewer seats as companies find efficiencies that reduce white collar headcount which reduces revenue for companies using seat-based pricing; 2) lower pricing power as AI-first upstarts offer reasonable alternatives that existing customers use to push back on annual price increases; and 3) lower new customer bookings again due to competition from AI-first upstarts. While this presents some potential headwinds, this may only apply where solutions can be easily ripped out and replaced with AI-first software. I believe there are still many examples of software that are not going to be easily replaced, and there are many parts of the software lifecycle that AI is likely to be less efficient at. I’m reviewing the wreckage in the software space, and there may be some additions there in time.
Existing Portfolio Activity
Trim: None
Sell: NVO
Add: DAVA
I sold Novo Nordisk (NVO) after a very brief period of ownership due to what appeared to be a disadvantageous competitive situation after further diligence. Eli Lilly (LLY) is much more highly valued than NVO, which initially attracted me to NVO as a value investor, but further investigation has revealed LLY has a superior product (for now). While NVO has been first to market with an oral GLP-1, it is likely only a matter of time before LLY brings a similar option to market. NVO may take an aggressive approach to pricing to access more customers and compete with LLY, but LLY’s advantage is likely to persist and I felt there were better alternatives for the capital.
I also added to the existing DAVA position during the quarter, while taking some tax losses on previously purchased shares. I continue to believe that DAVA is undervalued relative to peers who face similar AI threats as DAVA, such as EPAM. It has become clearer that “vibe coding” solutions are less likely to immediately displace solutions that require high levels of security and interconnectedness between users. With that in mind, there are still many applications that DAVA can help develop. I still believe DAVA’s share price performance relative to peers is mostly company-specific, and believe there is a significant chance of mean reversion if DAVA can demonstrate inflecting organic revenue growth. It however remains difficult to see the market disagree with my assessment, and this position is among my most fluid.
New Portfolio Activity
Bought: FND
Floor & Decor (FND) is a business I’ve wanted to own shares in for some time. They are the leading warehouse-style flooring store with higher in-stock inventory selection and lower prices than scaled competitors, and have been taking market share for years if not decades. In my opinion, they are following the Home Depot model to disrupt a profitable subcategory of home improvement. I still believe the current purchase price is not obviously cheap on near-term earnings, but the purchase price does reflect an attractive valuation on long-term margins.
The company’s current EBIT margins are about 30% below their long-term pre-COVID levels, and I believe EBIT margins should continue to scale towards the low-to-mid-teens as the company builds out its store base. Current sales per store are depressed by a post-COVID hangover and higher interest rates which have depressed existing home sales, a key catalyst for renovation activity. Higher sales per store will lead directly to higher store-level margins, which flow nicely through to EBIT margins. Analyst estimates have changed to assume very little same store sales (SSS) growth, after the company had very strong SSS on an annual basis until 2020.
I believe, with a more favorable existing home sales macro backdrop, that sales can grow at double digits with significant flow-through to the bottom line. It would not surprise me to look a couple years out and see the company generating $6.5B of sales at 7.0% net margins, which would mean the company is trading at 16x that admittedly uncertain (due to macro uncertainty) future earnings with nearly a decade of future store growth and comp growth. I see the company generating nearly $12 earnings per share 10 years from now, when it’s store growth plan should be essentially complete. It’s uncertain how the company would be valued given its higher cyclicality vs. Home Depot/Lowe’s, but the current purchase price likely provides low-to-mid double digit returns over the long-term.
Conclusion
AI is certainly changing the investment landscape, and I’m working very hard to stay on top of the changing dynamics. There are massive amounts of capital being raised to build or design new electricity capacity, data centers, chips, AI models, etc. There clearly appears to be an opportunity with AI to change how white collar work happens, but it’s not clear whether that will simply make existing workers more productive or create a more disruptive impact across the economy. The AI true believers certainly seem to believe in a more disruptive outcome, but I’ve seen enough can’t-miss buzzwords from Silicon Valley that didn’t really have the impact claimed that I am open-minded to a variety of paths AI could develop along, and I recommend you should too.
Best,
Mike Loeb
