←
The Fox1,710 posts
The Fox avatar

The Fox

@Alpha_Fox11

Breaking down businesses and stocks in charts. I run a business that sells on Amazon, so I read earnings like an operator.

◷ Joined August 2021▧ 536 Photos and videos
The Fox avatar
The Fox@Alpha_Fox11·Nov 12

Brookfield is hard to follow because the money moves in a circle. Here's how it works. BN is the parent. Everything on this map either belongs to it or sends it cash. One dollar can earn Brookfield money three times. Someone buys an annuity from Brookfield's insurer. The insurer invests that money in a Brookfield credit or infrastructure fund and keeps the spread. The asset manager, $BAM, charges that fund a fee. And when the fund sells its assets at a profit, Brookfield takes a cut of the gain. That cut is carry. BN owns 73% of BAM, which manages about $1 trillion and sends BN about $1.9B a year in dividends. It owns the insurer, which has close to $200B of assets, earned $511M in Q3 and folds into BN on Dec 1. It owns 45% of Renewable, 26% of Infrastructure and about 69% of Business Partners, which pay BN distributions and pay BAM to run them. And it owns the property group outright, which sends up about $630M a year. Carry is the piece the market trusts least. Investors in a fund get their money back plus a minimum return first. Above that, the manager keeps a share of the profit, typically around 20%. Nothing is paid until assets are sold, so the cash arrives in lumps, sometimes years after the profit was made. Most of Brookfield's carry belongs to BN. About $13B has already been earned on paper and is waiting on those sales. Brookfield values the whole stream at $34B, roughly $14 a share. Q3 turned $222M of it into cash, and the three year plan needs about $417M a quarter. That gap is the debate. Bears see a number on paper that never fully shows up. Bulls point to $53B of assets sold this year, which is the step that has to happen before carry gets paid. My read: the fee and insurance businesses already justify the stock. Strip carry out completely and the parts are still about $56 against $42. Carry is the upside, and the number to watch each quarter is realized carry against that $417M pace.

Post attachment
The Fox avatar
The Fox@Alpha_Fox11·11h

$SABR held its investor day yesterday. I went through the whole deck because Constellation owns about 12% of the company, and I wanted to understand what $CSU is betting on. Sabre is one of the big booking systems the travel industry runs on. About $270B of travel is booked through it a year, across 370M+ marketplace bookings. It connects to 400+ airlines and 2 million hotel properties, and it calls itself the largest corporate travel intermediary in the world. When a travel agency or a company's booking tool searches for a flight, there's a good chance Sabre is doing the work in the background. The last few years were a rebuild. Adjusted EBITDA went from about $100M in 2022 to about $610M expected this year. Almost all of its computing now runs in the cloud, and management says it has gained about a point of share every year since 2023. Leverage came down by about 7 turns over the same stretch. The plan they laid out runs to 2030. Revenue grows mid single digits a year, EBITDA grows high single digits, free cash flow adds up to more than $500M over the period, and leverage falls below 3.95x. Free cash flow is about negative $50M this year, and they expect it to turn positive in 2027. The growth is supposed to come from a few places. Payments is the fastest, with $26B+ of payment volume growing more than 25% a year. Hotel revenue has grown about 10% a year for three years. On the airline side, Sabre Mosaic is its new platform for airlines to sell and manage tickets, and three new airlines are moving to Sabre. There's also NDC, a newer airline booking standard. Airlines used it to push bookings onto their own channels, and Sabre is pitching itself as the way those bookings come back through agencies. AI got the most airtime, which makes sense because it's the main bear case. The worry is that people start booking trips through ChatGPT and the middlemen get cut out. Sabre's answer is that the AI tools will need it. It launched an MCP server more than a year ago, which lets AI agents plug straight into Sabre, and 80+ customers are using it in beta or live. It handles over a billion shopping requests a day. Their argument is that a chatbot can't price fares accurately on its own because of the infrastructure behind it, token costs, lag and security. I think that argument is reasonable, but I don't think the market is convinced yet. The stock fell about 7.5% the day before the event and another 2% on the day. The slide I found most interesting was on valuation. Sabre put itself at 7.8x 2026 EBITDA, against 12.0x for travel tech peers like Booking and Amadeus, and 17.9x for vertical software companies like Veeva and Tyler. Management is basically asking to be valued as software, which is how Constellation tends to look at businesses like this. The problem is the debt. Sabre has about $4B of net debt and expects about $494M of interest this year, against about $610M of EBITDA. The equity is only about 14% of the company's total value. That's where CSU comes in. Constellation paid about $86M for its 50 million shares, which are worth about $84M at $1.67. After Sabre adopted a poison pill in March, the two sides settled. Vela's CEO joined the board, and CSU agreed to stay under 15% of the stock, vote with the board and stay out of Sabre's debt. With a company this levered, buying the debt is often how an investor ends up in control. CSU gave that option up, so it's a regular shareholder waiting for the paydown. I ran the numbers. If Sabre hits its targets and keeps today's multiple, the equity would be worth about $3B versus about $670M now, and CSU's stake would go from about $84M to around $390M. If the multiple falls by one turn at today's debt level, about 90% of the equity value disappears. That math is rough and ignores possible dilution from Sabre's exchangeable notes. I'm curious where people think this ends up for CSU.

Post attachmentPost attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 17

$LMN Lumine closed two deals on July 1. On one of them it paid less than the assets were worth and booked a $4.9M gain. The other lost $6.8M in its first quarter under Lumine. Lumine is Constellation Software's telecom and media group. It owns 36 software businesses. About half are billing and customer systems for phone carriers, the stuff that runs when you sign up, change your plan or get your bill. Eight are TV and radio. The biggest of those is WideOrbit, which over 5,000 broadcast stations use to sell and bill their ads. The rest are apps that run on carrier networks and tools that keep those networks up. Nobody gets excited about billing software, but it's very hard to get rid of. Swapping out a carrier's billing system is a multi year project, and if it goes wrong millions of people get the wrong bill. So customers keep paying maintenance. 75% of Lumine's revenue ($218M last quarter) is maintenance and recurring, and deferred revenue is up to $143M from $95M at the start of the year. The part I didn't expect is that the old business has stopped shrinking. It's still choppy, but the last two quarters were positive. Through 2024 organic growth was negative every quarter, down as much as 12%. In Q3 it was up 3%, with maintenance and recurring up 6%. For a company that gets most of its growth from buying, that's the number I care about most. On the buying side, Lumine goes after the number one or two player in a small niche, doesn't mind if the business is losing money, and never sells. This year it spent $541M on three deals, compared to $20M in the first nine months of 2025: Synchronoss, $309M. A Nasdaq company it took private at $9 a share. About 1.8x sales. Imagine Communications, $137M. Broadcast video and ad software. About 1x sales. Quortex, $95M. A video business carved out of Synamedia. About 1.1x sales, and the one with the $4.9M gain. The market values Lumine itself at about 4.5x sales. So if they get these businesses running like the rest of Lumine, they bought revenue at 1x to 2x that the market prices at 4.5x. That's basically the Constellation model. The "if" is doing a lot of work though. EBITDA margin dropped from 36% to 31% in Q3 because of these deals, and management says Imagine and Quortex are only in the "beginning stages of strengthening." They don't merge anything. Every business keeps its name, its team and its own P&L, and Lumine gives them a playbook with 250+ best practices. I like how they finance it. Big deals get their own bank loan that's backed only by the business they bought. WideOrbit borrowed $175M in 2023 and has paid it down to $92M. Motive, which Lumine bought from Nokia last December, has its own $55M loan. If one of them goes bad, it doesn't drag the rest of Lumine down with it. Overall Lumine went from $142M of net cash at the end of 2025 to about $220M of net debt now, against roughly $200M a year of FCFA2S. Fine for now, but worth watching. As for Constellation, Lumine was started inside Volaris in 2014 by David Nyland, who still runs it, and was spun out in 2023 with the WideOrbit deal. Constellation owns about 61% and holds a single super voting share that always gets 50.1% of the vote. Lumine pays it a management fee ($802K last quarter) and reports the same FCFA2S metric. I like having Constellation in charge of capital allocation. The flip side is nobody is ever buying this company out, and minority holders don't get a say. At C$27.46 it's about US$5.0B, roughly 26x trailing FCFA2S. The stock is about half of where it was in July 2025 (C$54) even though Q3 revenue was up 55% year over year. If the new deals get back to last year's 36% EBITDA margin, my math says that's about $58M a year of extra EBITDA. The bear case is real. FCFA2S for the first nine months is down 14% because of interest, leases and weaker cash flow earlier in the year. Licenses and services are still shrinking. Telecom budgets are tight. And if AI makes this kind of software easier to replace, all that recurring revenue is less safe than it looks. What I'm watching: Imagine's margins over the next two quarters, whether organic growth stays above zero, and FCFA2S in 2027. Curious which of the three deals people think gets fixed first. Not investment advice. $CSU

Post attachmentPost attachmentPost attachmentPost attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 16

$CSU just bought Aspenware, the software a lot of ski resorts use to sell lift tickets and season passes online. The buyer is Andromeda, one of Constellation's nine operating groups. Andromeda has mostly been real estate, mortgage and lending software up to now, so ski resorts are a new vertical for them. The part I find interesting is who sold it. Alterra, the company behind the Ikon Pass, bought Aspenware with Aspen Skiing Co in 2022. Owning your vendor is great for you, but it's a tough sell to every other resort. Nobody wants to run their pricing and guest data through software a competitor owns. Under Constellation it's neutral again. It's grown a lot under Alterra too. 7M users and 2M transactions a year in 2022, now 12M+ users and 3M+ transactions across resorts in the US, Canada, Australia and New Zealand. Price wasn't disclosed, which is normal for CSU. One thing to keep an eye on: a price fixing lawsuit against Vail, Alterra, Boyne and Powdr points to three of them using Aspenware's dynamic pricing tool. Aspenware isn't a defendant and nothing has been decided. Pretty classic CSU deal. Mission critical software, sticky customers, and an owner that wasn't really a software company.

Post attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 16

Here is a 10 year model on $HONA. Discount it all back at 8.5% and I get about $194 a share today, so roughly 18% upside. Bear case $141, bull $245. Base case, the aftermarket grows 7% a year until suppliers catch up and margins get back to 26.5% by 2031. Even if the multiple drifts down from 15x to about 13x EBIT, that's around $446 by 2036, roughly 11% a year with the dividend. Over 10 years the bear case makes about 8% a year, bull 14%. What surprised me is how little is priced in at $165. You barely need margins to improve at all to justify today's price. That's without AI. If they get margin gains from AI in the plants and on the repair side, the base case goes to about $221 today, and even the bear case is worth about $164. On the debt, there's no paydown schedule. It's $16B of bonds due between 2028 and 2066 at around 5%. Their plan is to get to 2.5x leverage then buy back stock, and they already have $3.5B approved. If they wanted to they could pretty much wipe out the net debt by 2031. Management is targeting more than $6.5B of EBIT by 2030. I have $6.1B so I'm being conservative. Not a forecast, just my numbers.

Post attachmentPost attachmentPost attachmentPost attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 15

Seth Klarman's Baupost bought 1,132,000 shares of $HONA last quarter, a new position worth about US$175M at the end of September. Honeywell Aerospace has only traded on its own since June 29, when Honeywell spun it off. What it makes: the avionics, navigation and radio systems, auxiliary power units and engines, and control systems inside commercial jets, business jets, military aircraft and missiles. Three segments: Electronic Solutions, Engines and Power Systems, and Control Systems. The money is in the aftermarket. Last quarter US$2.1B of the US$4.6B in sales came from spare parts and repairs on planes already flying, against US$0.7B from new equipment. Once a Honeywell system is on an aircraft, it can earn parts and repair revenue for as long as the plane flies. The problem is supply, not demand. Backlog is US$19.3B, up 12%, and orders are up 9% over twelve months, but sales grew only 2%. Honeywell says a small number of mechanical suppliers are holding back output, and its own scorecard shows bearings as the worst of them. It is bringing about 700 parts in house and building a US$500M plant in Tempe that opens in late 2028. At US$165.01 on Friday, HONA is worth US$52.3B, about 21x the midpoint of its raised 2026 adjusted EPS guide of US$7.85 to US$8.00. It carries US$14.3B of net debt from the spin and just started a US$0.40 quarterly dividend. Why Klarman? Spinoffs are a classic hunting ground for value investors, because many holders of the parent sell shares they never chose to own. Baupost bought in the first full quarter after the spin, and it's now 3.5% of its reported US portfolio. My take: Klarman is buying the fix, not the current numbers. At 21x earnings with sales growing 2%, the stock needs the supply chain to catch up in 2027. The supplier scorecard Honeywell now publishes every quarter is the thing to watch. This is also the exact type of manufacturing business AI can bring margin expansion to: complex parts, a long list of suppliers to schedule around, and a repair business that is almost half of sales. I'd keep an eye on companies like this.

Post attachmentPost attachmentPost attachmentPost attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 15

$BAM earns fees on US$682B of other people's money and owns almost none of it. Pension funds, insurers, sovereign funds and wealthy families put up the capital. BAM picks the assets and charges a fee. That US$682B is up 13% in a year. Last quarter it produced US$1.62B of fees and US$864M of fee related earnings, a 54% margin. Distributable earnings were US$763M, up 15%. The mix has changed. Credit is now half of fee bearing capital, US$341B, after BAM bought the rest of Oaktree in July and kept adding insurance money from Brookfield Wealth Solutions. Credit fees grew 23% in a year, but credit pays less per dollar, so it's half the capital and 40% of the fees. The person to watch is Connor Teskey. He joined Brookfield in 2012, ran Brookfield Renewable, and became CEO of BAM in February at 38. Bruce Flatt called it the next step in a succession plan started four years ago, and Bloomberg has reported he's being lined up for Flatt's job at Brookfield Corporation. He built his name in renewables and is now pushing BAM into AI infrastructure, including a Korea deal with NVIDIA and NAVER. BAM or $BN? BN owns 74% of BAM, plus all the carried interest on older funds and a third of it on new ones. BAM is the clean fee stream. BN is that stake plus insurance, real assets and more of the carry. Some growth is already signed: US$148B of commitments aren't invested yet, and BAM says US$68B of that will add about US$680M a year of fees. At Friday's close of US$51.53, BAM trades at about 28x distributable earnings and yields 3.9%. The US$0.5025 quarterly dividend is above last quarter's US$0.47 of earnings a share, so fees need to keep growing in the mid teens. My take: the business is strong and the succession is clear. At 28x, you're paying for Teskey to keep it compounding at 15%, with little room for a slow fundraising year.

Post attachmentPost attachmentPost attachmentPost attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 14

Two adjustments I made, and what has to be true. Asseco paid Topicus €59.0M in Q2 2026 against €7.7M a year earlier. Per share that was PLN 13.05 against PLN 3.94 in 2025. Topicus owns 23.1% of Asseco. I kept last year's dividend rate and took the rest out, which cuts trailing FCFA2S from €250M to about €225M. And 38% of the cash Topicus's businesses generate goes to minority holders, much of it to the Joday Group through Topicus Coop, so I only count what's left for Topicus shareholders. The grid shows the reinvestment needed under other assumptions. If deals only earn 10%, it's 74%. If organic slows to 3% and deals earn 10%, even reinvesting everything doesn't get there.

Post attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 14

What each pace is worth today on the same assumptions. 36%, priced in: C$100.50, about C$192 a share in 2036. 38%, this year's pace: about C$103. Half: about C$120, C$251 in 2036. Three quarters: about C$167. All of it: about C$241. Topicus closed the Softland acquisition on November 6 and didn't disclose the price. How much it put to work in Q4 is the number I'll watch when it reports.

Post attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 14

$TOI Topicus reverse DCF. No exit multiple this time: a 10% required return, and growth built from organic growth plus how much cash gets reinvested and what it earns. At C$100 the stock trades at about 24x trailing free cash flow to shareholders, once you take out the extra Asseco dividend (more below). With 4% organic growth and 15% returns on deals, roughly what $CSU has earned, the price needs Topicus to reinvest about 36% of its free cash flow for the next ten years. In the first nine months of 2026 it reinvested 38%. In 2025 it reinvested 191%, mostly the Asseco stake, and borrowed to do it. So the price assumes this year's slow pace continues. Any pickup is upside.

Post attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 13

$CSU is one ticker, but it is really three companies and a put option. Constellation has nine operating groups. Seven are held directly: Volaris, Harris, Jonas, Perseus, Vela, Modaxo and Andromeda. Together they did US$2.71B of revenue in Q3 with 1% organic growth. Harris includes Altera, the hospital software business bought from Allscripts, at US$144M a quarter. The other two trade on their own. CSU owns 31.3% of $TOI Topicus but controls it with one super voting share worth 50.1% of the votes. $LMN Lumine works the same way, held through Volaris: 61.4% of the economics, 50.1% of the votes. That matters when you read the numbers. CSU reports all US$799M of their Q3 revenue, 23% of the total, but only part of their cash belongs to CSU shareholders. Both also grew faster organically than the rest of CSU, 5% and 3%. Then the put option. The Joday Group, CSU's partner in TSS since 2014, owns 29.4% of Topicus Coop and can sell those units to CSU. That liability is US$1.3B and moves with Topicus's value, including its roughly 25% stake in Asseco Poland. It is the line behind Q3's US$117M non cash charge. At market, CSU's Topicus and Lumine stakes are worth about US$5.8B, around 11% of CSU. Close to 90% of what you pay for is the business it holds directly. Thanks @adrivalue for flagging the new groups.

Post attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 13

My read: a solid quarter dressed up as a great one. Volume, retention and cash generation are real, and so is the cost discipline. But the record gross profit needed FX help, the take rate has not found a floor, and Q4 has to beat Q3 by about 15% on operating profit to hit the guide. The valuation already assumes a lot of that compression continues, which is why my base case sits above the price. Before paying up I would want two quarters of gross profit growth without help from the Others line, and Mexico margins back up. Next to watch: the Q4 take rate, Q4 opex, and the 2027 guide in March.

The Fox avatar
The Fox@Alpha_Fox11·Nov 13

At Thursday's $13.70 close the market value is about $3.9B. With about $373M of net corporate cash, that is roughly 14x trailing operating profit, 18x trailing earnings and a 6% free cash flow yield. At 12x operating profit, the price needs about $297M of operating profit in 2027, only about 6% above the low end of this year's guide. My scenarios: bear $11.13, base $15.22, bull $19.44.

Post attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 13

Guidance: TPV growth at the top of 60% to 70%, gross profit near the middle of 25% to 30%, and operating profit toward the low end of 27.5% to 32.5%. That implies Q4 gross profit of about $136M and operating profit of about $87M, up roughly 15% from Q3. Hitting both means opex of about $48M, 13% below Q3. Either Q4 gross profit beats the midpoint, which seasonality makes possible, or costs fall again. Management also said FX could swing Q4 gross profit growth by about 6 points either way.

Post attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 13

Customer quality is strong. Net revenue retention was 140%, and of the $116M of revenue added in a year, $112M came from existing merchants and only $4M from new ones. The flip side is concentration. The top 10 merchants were 62% of revenue. Those are the merchants with the leverage to push for lower pricing tiers, which is what shows up in the take rate.

Post attachment
The Fox avatar
The Fox@Alpha_Fox11·Nov 13

dLocal has bought back about 9.1M shares for $120M under its $300M program, around $13.20 a share, and the share count is down to 287.4M from 294.9M at year end. Over the same stretch General Atlantic, the largest holder, filed in August to sell 9.84M shares worth about $137M, roughly $13.91 each. It still owns 21.4% of the Class A shares. The buyback has absorbed a lot of that supply.

Post attachment