Brass Tacks Capital

Portfolio, End of Q2 2025 $META $GOOG / $GOOGL $SHW $V $AMZN $MSFT $AMT $HD $CHDRY $UNP $ICE $SPGI $ANCTF $ECL $PAC $WSO $ALC $DG $FSV $DIS $LRLCY $TXN $ZTS $OTIS $WCN $ASAZY

$MSFT looks glaringly cheap for a large cap. If you zero out all of OpenAI's revenue contribution and you're ~24x trailing earnings for a utillity which even without AI would grow revenue double digits for a long time.

Evercore on recent $V and $MA underperformance, noting that they are increasingly being valued on terminal growth rather than on NT execution:

I keep coming back to how good this whole “financial tollbooth space” is. Businesses companies like $SPGI, $MCO, $MSCI, and $FICO all embedded in the system and take a small cut every time money moves or decisions get made. Since they are asset light and don’t require much capital, a lot of the revenue turns into cash and it is a great economic model. $SPGI and $MCO are great because they’re tied to credit markets. If companies want to issue debt or investors want to understand risk, they’re part of that process. It’s hard to replace them once they’re embedded. $FICO is similar in terms of quality. It has a very strong position and great economics, but it’s closer to the consumer so when it pushes pricing too hard, attract unwanted attention. It’s still a great business, just a bit more exposed. $CME is a different kind of tollbooth. It doesn’t charge for data or ratings, it charges for activity. Every time someone trades futures or hedges risk, $CME takes a cut. The beautiful part is that volatility actually helps them. When markets get messy, volumes go up, and so does revenue. So instead of being hurt by uncertainty, they benefit from it. There are more businesses in this category like $ICE, $LESG, $NDAQ, etc. But $MSCI is the one that stands out to me. On paper, the numbers might look similar, but the type of pricing power is different. $MSCI doesn’t charge individuals or even single institutions in a meaningful way. Its fees are spread across trillions of dollars tied to benchmarks, so almost no one feels the increase directly. That makes its pricing power less noticeable, and because of that it’s better and more durable. What’s interesting is the meaningful insider buying stock in the open market at prices not far from where it trades today. That’s not something you see often in already well understood, high quality businesses. So I really do think this financial tollbooth space is a great place to invest. $SPGI, $MCO, $CME, and $FICO are all excellent. But even with all that, I still come back to $MSCI as my favorite. The pricing power is just more elite. Fees are spread across trillions tied to benchmarks, so no one really feels them, and flows come automatically through passive investing and asset allocation. It’s simple, embedded, and very extremely hard to disrupt. 🌹

Visa and MasterCard processed ~$21 trillion (with a T) across the last nine months of 2025. That’s roughly $54 million per MINUTE. So the entire AI agent volume over those nine months is processed by $V + $MA every ~50 seconds.

$META check-in: - 20x GAAP PE - Growing revenue +30% yoy (engagement, CPM, then ad-load) [double the rate of $GOOGL on half the revenue base with a lower multiple] - User and engagement growth across all platforms - Reels engagement growing +30% - Threads and WhatsApp top 10 most downloaded apps (and WhatsApp #1 messaging platform globally) and both completely unmonetized - OpenAI pulling back on consumer (Meta won with impossible to replicate distribution) - All the costs of non-core AI in the financials with little of the benefit; 1B in AI DAUs and new team in place 8 months closing capability gap with frontier labs - 20% RIF (or more) likely coming soon worth likely HSD to mid/high teens % on EPS runrate [and most know they could probably double this and see no negative impact on operations] - Reality labs worth a few turns on the multiple and both masks underlying profitability while also providing substantial platform optionality from AR / AI agents - Building ahead on compute infra which means they will have by far the most 1P inference infra of anyone (and all of it will be pointed at the consumer)

I suddenly happen to be interested in French politics

Here's the current state of the Durable Value Creators investable universe. Spent some time recently narrowing down the focus. Primary themes include picks & shovels of secular trends, recurring revenue, razor-razorblade model, mission-critical, and/or relatively asset-light.

Inflation running hotter than consensus, not better place to be than $MA & $V. Both off ~ -15% off the highs, growing forward EPS & topline DDs. No brainer!

Both Mastercard and Visa are now firmly below their 10-year average price-free cash flow multiples, and heading towards the lower band. $MA $V

I agree w/ the thesis that payment for labor will become payment for tasks/agent services/software I am buying V/MA here Massive TAM expansion

The below excerpt from Pershing Square's 2025 annual letter sums up the opportunity in $META nicely I think. ~22x earnings for one of the best businesses in the world seems reasonable to me. #page=9" target="_blank" rel="nofollow noopener">assets.pershingsquareholdings.com/wp-content/upl…

The fact the no one realizes how high quality of a business $META actually is is exactly why it is so cheap and why it’s such a good risk/reward imo. People will figure this out eventually and the stock will have rerated to a much much higher multiple - people will only explain this in hindsight with price as their guide.

Added at $306 $V

1/ 80% to 90% of high school graduates in East Asia now suffer from myopia (nearsightedness). It's an epidemic. And there are dozens of companies benefitting from the shift.

$ALC

$msft is 2% away from the 200 week, it hasn’t closed below that level since 2013.

reminder that agentic commerce is a fraud nightmare and should be massively TAM expansionary for the payment rails $MA $V

$META and $GOOG operating profit trajectory is a thing of beauty. Happy to hold them as my top 2.

Better way to express the “water investment thesis” than Ecolab? $ECL CEO: every part of the AI value chain depends on water. Fabs that make the chips, power plants that fuel the chips, and data centers that run and cool them.

For example: - $AMZN reported great numbers in 4Q w/ accelerating growth in AWS, and its role providing infrastructure to support Anthropic (which is growing explosively and accelerated in Jan/Feb) should cause AWS growth to incrementally accelerate through 2026. - Why is the stock down? A 1Q guidance miss on margins (in part due to expensed satellite launch cost that will be capitalized in 2H) and a larger than expected capex guide spooked investors and caused 2027 EPS estimates to fall by ~2%. The stock is now -22% from Nov 3rd 2025 highs. - What is the market missing? AWS grew ~24% in 4Q25 and estimates for 26/27 are for continued ~24% growth. It strikes me as quite conservative to assume growth does not continue to accelerate even as AWS grows capex ~53% yoy in ‘26 (AI capex more like +60-70%) and with Anthropic adding ~$4-5B in NNARR in Jan vs ~$8B added in all of 2026. - What will change investors minds? If AWS reports growth in the high 20s in 2H26 it would not surprise me. AWS is much higher margin than retail, so a ~2% increase to company revenue coming from AWS could drive a ~MSD increase in ‘27 cons EPS estimates to ~$10/sh. Walmart is currently trading at 50x EPS, so I think it’s not crazy that by YE26 AMZN could trade @ 25-30x $10 (also its pre Feb’26 mult), yielding a $250-300 stock at year end or a +25-50% return from today. PS: Satellite launch cost moving to capex from opex, lapping tariff and NBA cost headwinds, and ongoing mix shift to ads and 3P services should provide more upside to EBI.T.