Pinned straw:
Interesting that they have decided not tp provide third quarter results. This is a departure from their normal reporting program. I suspect recent market weakness has priced in a poor 3rd quarter and is not looking for a huge uptick in the final dividend.
Having said that I do think the previous commentary on the tailwinds that exist such as PC refresh cycles and evolving AI use cases will hold the company on good stead.
Held IRL.
I concur with much of the analysis in this thread. I have been adding to my position in real life and added a little bit here on Strawman.
It's a risky play but I do think it has the potential to have a significant payoff in the medium term.
@Slomo - I agree with the comments of others - this is a great piece and captures many of the ideas in my own thesis. Without repeating the comments of @Solvetheriddle @Karmast and @thunderhead , I do want to say that I am always comforted when @Rick comes along with McNiven, and underpins the low end of my valuation!
First up, $DDR is not my favourite kind of distributor for reasons you've touched on. Its large market position and respectable margins in its peer group does create an incumbency risk. I do believe that, like the Autodesk case, it is becoming easier for suppliers to go direct. Online models and 4PL fulfillment capabiliites in hardware and software continue to advance, enabling owners of products and services to acquire and support customers directly at ever lower costs. So Autodesk is a flag to keep an eye on, as it is a potential systemic risk. (I prefer attackers with proven economics at low scale, who can then grow into highly fragemented markets, with favourable incremental economics. But let's not get too academic about this - an opportunity is an opportunity!)
In terms of competition, while $DDR's margins are a positive signal to encourage new entrants, they are also to some extent a moat. $DDR can use its superiour purchasing power and ability to selectively flex margins effectively to protect its market position.
My thesis has two elements - and I think I probably wouldn't have invested in $DDR without both (even though I acknowledge the views of several respected StrawPeople to the contrary.)
The Base Case
First, there is no doubt we are going through a cyclical trough as has been commented on widely. Both macro and sector-specific. Are we through that? I'm not sure. But I agree with your remark that by the time it is clear we are through it, the market will move quite quickly, and a decent portion of the upside will be gone. I think its worth taking the 6-month or 12-month timing risk on that. (Recognising a recent exchange with @Rick - I've placed half my chips on this, saving 50% for when momentum is confirmed. Learning in action!)
The normal cycle in that sector, has two elements. The broader economic cycle. Easing interest rates will flow through to a recovery in business investment. And we are also entering a likekly 4-5 year business IT replacement cycle in hardware, as well as the software retirements you spoke about.
These elements likely offer the potential for anything like 15% to 30% annual reutrns, including dividends, over the next couple of years, with a limited downside (maybe 5%-10% net of dividends), unless these is evidence of a secular decline trend (Autodesk+++).
The Upside
AI. Before you all groan and switch off, I believe that over a 2 - 5 horizon the AI phenomenon can turbocharge both the software and hardware replacement cycle.
I've been doing some lightweight research on the sector. The hyperscalers are already investing $100 billion 2023-2024 in AI infrastructure, and there are estimates this can go to $1 trillion in 2024-2027, and I yesterday listened to a GS podcast which referred to $1.5 trillion in capex (although I don't recall the timeframe). This has fundamentally changed the business models of the hyperscalers from capex-light growth to capital intensive, although in truth the change began when they started going heavy on cloud infrastructure. And of course, we all know that AI-processing is an order of magnitude higher in energy intensity, so we're hearing stories about hyperscalers investing in nuclear power plants. Talk about the polar extreme of capital light! What a transition - a whole industry flipping from capital-light to capital-heavy in 3-5 years. Amazing. Now these are not pipe dreams. It's actually happening today.
So the sector is frantically working to develop AI-chips for phone, laptops, PCs and servers. Again, its happening, The architectures are developed and continuing to be developed further, the factories are tooled up, the first models of the chips are rolling off the production lines and into products.
Yesterday, I finally joined the dots (I'm a slow learner). The hyperscalers NEED the AI-chip embedded products to help shift as much of the energy consumption from them to the customer. I have to do more research on this, but it seems that they can shift at the low end 10% and as much as 30% onto the end user. It may be more - I have to look more deeply in to this. However, it is already clear that power is going to be the big opex cost for the hyperscalers, and so reducing this is fundamental to their economics, and therefore to their returns on investment in infrastructure. Through chip design they will push as much processing on to users as they can.
Of course, its not just about power. User experience is a factor as well. Putting more parallel-processing on the end device reduces latency and accelerates image and video processing as well as natural language processing. In jurisdictions like the EU with stringent data privacy regulations, it can also help keep more user-sensitive data on the device, aiding privacy and cyber security.
So while I don't know where along the value chain, how large, and when the returns are going to arrive for the huge investment in AI (and I don't think anybody knows), I'm pretty confident that AI is going to turbocharge the next phase of end customer investments in hardware and software.
Just think about the technology we already have. Most business users don't use a fraction of the capability and processing power in their laptops and desktops. And yet, if a business mandates "Core-i7" is the company standard archiecture, everyone gets it. (I saw proof of this a couple of years ago when I was involved in a charity effort to source cheap refurbished laptops. We ended up buying a big stack of cut-price refurbished laptops from a government department - they were three years old, and at time of purchase all had the latest generation Core-i7 processors, GPUs and expanded RAM specification, even though I am sure many of them were mainly used for email and playing Solitaire!) I think we are going to see the same phenomenon with the AI-fleet.
These two elements of the thesis make $DDR a compelling proposition for me. I don't know how strongly the AI Upside tailwind is going to blow, but for me it makes the potential returns for $DDR strongly assymetric over the next 2-5 years. It's probably not a super long-term hold for me. But it look like a "picks and shovels" beneficiary of the investment everyone else is making in AI. And $DDR doesn't need to invest an incremental dime to play, it just has to keep on doing what it's already doing! How good is that.
Disc: Held in RL (5%) and SM - prepared to go significantly bigger, subject to next HY results
@Slomo great effort on this one. I think most of the issues are timing and cyclical, as you point out. I am a bit concerned if the market evolves away from then Vlad did say AI is a harder and longer cycle sale. We shall see, that’s my biggest risk. It’s not an outstanding business but reasonable quality, a poor man’s tech play lol
held may buy more down here