All Activity

Thank you!

Of course, as you may know I agree with all of this. My only response would be from these four levers, the only lever not fully in their control is interest rates. And my point was if you take the others as given i.e great capital allocation and alignment, the variable macro i.e interest rates is also in a great spot i.e 5% locked in for ~3years with rates likely to stay there or higher rather than going lower. So the downside risk today is very low.

My aim is next 6-12 Qs. I think SNDK will do $400+ in CY27 EPS. From that perspective this is very cheap, don't even need a re-rate. Then you have HBF as well. Kimi K3 comes with a new KV cache optimization layer that shrinks cache by 75%. It is now 4x cheaper to just let cache stay because if you're not doing that then whenever the user comes back to the chat the labs will have to re-compute everything on the GPUs. I have kept trimming SNDK. Started the position in Tegridy when stock was $400. It kept ballooning and becoming a bigger % of the book. Cut the proceeds from it to fund LQDA buying when the stock went down as well. Whats in the fund now is basically house money, or even less than that. NVDA ICMS architecture specs out roughly ~16TB of KV-cache capacity per GPU, ~150TB of context memory pool per rack, built from 100TB SSD modules, scaling up to a 10PB ICMS cluster serving an NVL72 rack-scale system. The Rubin generation that actually carries BlueField-4/ICMS enters rack mass production end of 3Q26, with ~5,000-7,000 VR200 racks in 2H26 and the real volume ramp in 2027 according the street. The ICMS central case of 85-90EB would consume 60-75% of the entire industry's annual bit growth by itself. Then you add non-NVDA demand and we'd still be constrained. I think r/r for SNDK very good under $1,600/sh.

Like most others here I am also surprised at the relatively low valuation Fairfax trades at compared to its fundamental performance. Mr. Market can be funny, unpredictable and most importantly wrong for long periods. That's what makes investing hard but also interestingly also what makes it easy, ie these opportunities exist not for a week or a month but can exist for a year+. Based on my experience it has started to happen more often recently and for longer durations as a lot of value investors complain (Terry Smith is a recent example, David Einhorn a less recent example), maybe because of increase of passive and short termism from quant funds. In my opinion, the only solution is you invest in companies that have great capital allocation, i.e if the share price is depressed and Mr. Market does not give the credit due they use their cash flow to take advantage of the dislocation through buybacks as Fairfax does. You get EPS compounding regarding of what Mr. Market or the multiple does. The question I get most when I talk about Fairfax is why does it trade at 8x earnings - something must be wrong with it. Investors try to explain and rationalise Mr. Markets moods. As opposed to a first principle analysis of what a business trading with 15-20% sustainable earnings power (and 15-20% ROE) over the next 5 years should trade at.

Well yes, that was the issue with hockey too. Dead puck era was also known as the clutch age grab era. Too much obstruction, tactical interference, offsides and two line passing bans stunted gameplay; essentially the defensive model used the rulebook to perfection. Scoring was the lowest ever, the game to many was boring…and eventually they changed the rules, fixed the offside bullshit, and opened the game up to high speed, high skill play. Ironically, the Russian style of hockey that was much maligned through the 80/90s. Soccer could certainly benefit from this, especially changing the stupid offsides ruling even partially…if 80% of the players are on one side of the field there’s zero point to calling offsides, let alone the disgrace we saw in the Croatia game where everyone was in the box. Make penalties for embellishment severe.

you are not nervous about "peak earnings" for SNDK in 2027? When supply catches up to demand then suddenly the price of memory will plummet and at 7 times 2027 earnings it could still be expensive if it earns way less in 2028 and beyond. I am willing to bet the capex cycle continues longer than people expect but "peak earnings" is one factor kind of making me nervous about SNDK. Or you will hold out for 3-4 quarters of great earnings, wait for the stock to get re-rated, then get out?

I’m not sure this makes sense. Almost all innovation has been private at inception for a long long time. It still eventually increases productivity and incomes broadly and usually leads to lower prices. AI is essentially free right now. If people are not getting benefits, that is on them or on AI being over-hyped currently. Giant flat screen TVs are now less than $100. ICE engines now get twice the miles per gallon that they did 20 years ago and gas is roughly the same price (without Iran war) due to the shale drilling innovations. Those things make everything more productive and cheaper for everyone. I mean the internet has made everything cheaper and more productive in millions and millions of ways. AI will eventually be the same. There is just no way the benefits all go to Zuck. Creative destruction is alive and well in the USA.

"The agreement makes Kraft Heinz the exclusive provider of some condiments, macaroni and cheese and cream cheese at Disney’s North American parks and resorts, and on the Disney Cruise Line. The deal enables Kraft Heinz to use Disney characters and stories across 10 of the food company’s brands, including Heinz, Philadelphia and Kraft Mac & Cheese. It also extends to Disney’s studios and streaming platforms, where Kraft Heinz will fund content co-created by the two entities." https://www.wsj.com/business/media/kraft-heinz-strikes-deal-with-disney-to-supply-resorts-and-tap-characters-d24a8b15

+1 I actually thought the spain / argentina game was great too (aside from argentina playing dirty). Watching a good team (spain) consistently move the ball, find space is great to watch

Too much dirty stuff is actually allowed, that is the problem. Defenders get away with so much, it stops great football. Have a look at indoor football where this stuff isn't allowed, just a gorgeous sport.

Thank you. Great letter and returns.

Agree in principle and even moreso in the historical context of the 1900s, but I think this argument is getting tougher for people to buy. Take Artificial intelligence as the latest great technological achievement: How many billions has AI created for Zuck's net worth? How many people has AI lifted out of poverty? Not meant as a scientific analysis but it gets to the point: the benefits of scientific, technological, and economic growth are increasingly being privatized.

Added to HONA, AMRZ, ASR, LILAP. Rebought some BOL.PA

+1

I think this framing is problematic. The economy is not a zero-sum game. You can improve the living standards for poor people without getting rid of the rich people.

I think it will take a hard market not to sell materially below intrinsic value. Under the current market structure I think the multiple range will oscillate lower during soft markets making higher lows vs the last cycle and higher highs under subsequent hard markets. I used to think they would issue shares again if the stock is near intrinsic value and they might if the right opportunity came along but I also think they want to reduce the share count enough that employees and the family have enough votes after Prem’s super voting shares are reduced to single voting to block any hostile bids.

Lol this is hilarious. Maybe you just don't like soccer. If you don't see the beauty in the game that's fine and. I for one think it is spectacular because I have played for my entire life and I know the intricacies and what good really is. With admitted bias I thought the Spain v France game was incredible. To shut down the worlds best offence shows that Spain is playing an elevated game. I don't think there is any comparison at this point and I think this win will push back the flow towards power and speed over skill and football IQ. I will say the past 15 years of sub par leagues full of shit players adopting the possession tactic has made the game slower and less exciting. Watching Pedri or Dani Olmo cycle the ball with their D like it is on a string is exciting. Players from Toronto FC trying to keep possession in their half and involve their goalie only to get scored on is pretty awful. The joke when that happens is Pep Guardiola has blood on his hands. Lower tier teams in the Premier league struggle with this as well. The game is not perfect and a much stricter policy on embellishment that combines financial penalties, in game consequences and public shame would go a long way to help.

It's been that way for 20yrs. And for a period during that time(not recently) it was also deserved. What is a little surprising is that after the last 5yrs of results, insurance operations, investment performance and interest rates reset, that's still the case. Sometimes these imbalances do take a while to correct. I was just comparing Fairfax with WR Berkeley across a range of metrics last night and my conclusion was that it was about 1/2 to 60% as cheap. Thats insane as it was on average as good as or better on qualitative measures as well. I think the reasons have to do with what Viking and other board members keep saying. They're not splashy or promotional, they're listed in Canada, they're not on the radar of a lot of N.A. institutional investors, their books are more complex than the average insurer as are their investments. They under report their BV if anything. However, they also have a lot of consistent and long term investors perhaps as much as 60-65% by some estimates. And they keep chipping away at the balance of weak hands with their share buybacks. Yes we are surprised that the luxury of undervaluation is lasting this long, but the longer the merrier if you believe in their long term success. This gap will eventually close as remaining holders demand higher prices for their shares. They have stable and almost predictable streams of cash coming in over the next 3yrs at a minimum. I think most of us feel a major cat. year is about a break even. Outside of that, share buybacks will likely continue at these valuations. The only reason this might be a little frustrating is if someone is nearing a need for cash, a wedding, a house purchase etc and have to liquidate some holdings out of necessity. Finally if that day does come where shares are no longer selling materially below IV, a couple of things happen. They will have other excellent uses of cash, like buying minority shares in subs, New private equity deals via KW, non insurance acquisitions, debt paydowns, Indian infrastructure, and so on. Lastly if the shares were ever trading significantly above IV, I believe they would consider acquisitions using shares also. I also like how they make acquisitions with debt to reduce the overall cost and keep the holding company as a backbone to support them while reducing leverage. Overall it's a good set up and the business model is the engine that is driving this train.

A little Fairfax, MSCI and Constellation Software.

BWET has been on a rampage, if crew wages go up then this will get truly nutty.

I don’t mind. I’ve been adding every paycheck. I’m just surprised the opportunity still exists, but then when I listen to pitches like the recent TIP podcast I can understand why it does.

Now we know it will be funded through debt issuance. S&P updated outlook to Negative due to risk of not restoring leverage to 5.0x in next 12 months. Some information below suggests Sleep Country had noticeably improved profitably in 2025. https://www.spglobal.com/ratings/en/regulatory/article/-/view/type/HTML/id/3597931 On July 21, 2026, Sleep Country Canada Inc. announced its plans to acquire Sleep Number Corp. for a total purchase price of $725million. The company plans to issue a $270 million term loan A and $400 million senior unsecured notes to fund the transaction. We assigned a 'BB-' issuer credit rating to Sleep Country Canada Inc. We also assigned a 'BB-' issue-level rating to its proposed $400 million senior unsecured notes. The recovery rating is '4', indicating our expectation for average recovery (30%-50%; rounded recovery: 40%). We estimate pro forma leverage of about 7.4x for the 12 months ended March 30, 2026, which is above our 5x downgrade threshold for the 'BB-' rating. The negative outlook reflects the likelihood of a downgrade if we believe Sleep Country is unlikely to restore its S&P Global Ratings-adjusted debt to EBITDA to 5x by mid- 2027 due to weaker consumer sentiment or a slower-than-expected turnaround at Sleep Number. Management has identified at least $67 million in cost synergies to be implemented through 2027, primarily through lease renegotiations, supply chain optimizations, and procurement efficiency, specifically the replacement of suppliers to reduce input costs. We expect synergies from store rationalization and lease renegotiations (completed via the Chapter 11 process) to be realized immediately post-closing. Procurement synergies are also expected to materialize shortly after closing, given the short lead time for inventory orders. Nonetheless, Sleep Number’s margin profile will remain 14% compared with Sleep Country’s 20% EBITDA margins in 2027. Consequently, we forecast Sleep Country’s EBITDA margin will decline to 16% in 2027, reflecting the dilutive impact of Sleep Number’s lower-margin profile. Sleep Number has incurred $236 million in total restructuring costs since 2023, which we do not expect to recur going forward. We expect the company to prioritize its strong free operating cash flow (FOCF) for debt repayment. Sleep Country ‘s FOCF has improved due to higher profitability. It reported FOCF of about $69 million in 2025, compared with $30 million for the same prior-year period. We expect FOCF will be depressed in the first year post acquisition. Absent any further restructuring and transaction related costs, we forecast FOCF will rebound to about $115 million in 2027. We also expect capital expenditure (capex) to decline to 2.5% of revenue going forward compared with 3.5% before the acquisition. We forecast it will use FOCF toward its deleveraging plan, prioritizing its term loan.

When Fairfax makes a really big investment, do you want them to buy high or buy low? I much prefer they buy low. And the lower the better. Fairfax's low share price of the past year has been a massive gift for long term shareholders. We should be celebrating our good fortune. Volatility in the stock market is a wonderful thing. It provides opportunity. Fairfax's stock is crazy cheap. And Fairfax is feasting.

Getting difficult to keep tankers crewed in the SOH. https://gcaptain.com/shipowner-offers-seafarers-six-months-pay-to-sail-through-strait-of-hormuz/ " Sinokor Group, the world’s largest owner of supertankers, offered six months extra salary if they make a return voyage, which it said would take about a month in total. [Other] shipping firms were adding as much as 60 days of salary to a 30-day contract". The highest-paid person on most ships is the captain, who has the ultimate say on navigational decisions. They earn as much as $15,000 a month on oil tankers, according to officials at two shipping companies. A junior sailor, known as a rating, might earn about $1,500 a month in normal times. They have the right to ask to leave a vessel and be replaced if they don’t want to sail into the danger zone. “We have heard stories of a large number of crew members getting off, but they are able to find people who are willing to go.” Junior ratings are not rich people, and typically have families reliant upon their continuing income to survive .... things have to be extremely bad for them to give it up. One of these under-crewed tankers gets attacked, it's abandoned at sea .... salvage for the picking All this before the Houthi threats on Red Sea tankers, and the closer proximity of Somali pirates looking for salvage. SD