A Look at IBM
Board: Berkshire Hathaway
So, full disclosure, I'm not a BRK shareholder. I find it an interesting stock, but I've never pulled the trigger. I do consider myself an expert on IBM, however. The one unifying fact about my career for 15 years (over a half dozen companies) was that I worked for companies that kicked IBM's butt. That was until the company I was working for was kicking IBM's butt so badly that IBM decided that they needed to acquire us to take us out of the marketplace. :-) So then I worked for the "enemy" for several years. In a bunch of roles, from technical expert through mid management. (I left IBM over six months ago, so there is no insider information or conflicts of interest here. I own no IBM stock.)
Anyway, one reason that I can excuse Mr Buffett not being able to rattle off so many details about IBM that it can about KO or other stocks is that IBM is extremely diversified and complicated. If you asked me what kind of moat IBM has (or how much it costs to acquire a customer), I'd have to counter with what kind of customer? Hardware? Software? Systems? Consulting? A customer acquired through an acquisition? A multi-brand "end to end" customer? Or a single product "best of breed" customer. If there is one defining characteristic about IBM is that its overall size compared to its competitors.
The short answer to the question is that IBM has two moats. The first moat is what I mentioned before, it is bigger than everyone else. (Oracle is close, but beyond that there is no one.) If you are a CIO at a Fortune 500 and you are sick of trying to get all of your small tech vendors to play nice with each other, give IBM a call. When IBM acquires a small company all of those CIO's will suddenly give that new product a try because they know that IBM will "just make it work" with the rest of their IBM software. In this way, IBM has a boat just like InBev's moat. Their products may be inferior much of the time, but it doesn't matter because their distribution channel is superior.
Another interesting moat that IBM has is its international channel. This is somewhat related to the first moat, but outside the US, IBM has a distinct advantage. Most hardware, software, and systems vendors don't have the size to be truly international. Having consultants across the world, support people who speak native languages, translators to help port software documentation, lawyers that understand all of the regulatory complexity, shell companies to manage all of the profits, and marketing and sales people across the globe really helps IBM provide unique value. A big part of how IBM justifies all of the acquisitions it makes is the fact that it knows that it will be able to rapidly increase the sales of its acquisitions by bringing those innovative products using its international distribution channel in ways inconceivable to smaller companies.
So, with that said, is IBM a good investment? I'm not so sure. I have some of the same concerns that the linked article has. IBM is stagnating revenue wise. Even with all of the acquisitions and all of the advantages I mention above, IBM is shrinking in terms of how much product it ships. All of its earnings growth has come through expanding margin (largely through reducing costs). IBM has declared an earnings goal of $20/per share for 2015, and as an employee you got the feeling that IBM was going to meet that goal no matter what it took. No matter how much "into the bone" the cuts in expenses might be, IBM was going to make that goal.
And to respond to mungofitch (no offense, I think you are a smart guy), IBM is absolutely gouging its customers. Everyone knows it. IBM knows it. Customers know it. IBM's goal is to be come so central to their customer's business that they can charge whatever they want. IBM's customers goal is to become such a large customer of IBM's that they can demand outrageous discounts to combat this (an/or to get rid of IBM). IBM's first move when they acquire a company is to quadruple the price. Especially knowing how their largest customers will demand 90% discounts. Price is not how IBM wants to complete.
So, from the short term, I think you can look at IBM and say "in 2015 it will be earning $20 a share". Multiply by a reasonable P/E and there isn't a whole lot of downside to the stock. Especially as a "blue chip".
But Mr Buffett doesn't tend to be a short term investor. And long term, where goes IBM? I can't look at a single product IBM makes and say "this product is going to be more valuable two years from now than it is now". And IBM makes a LOT of products. So, long term, IBM is not where I want my money to be. And it always surprised me that IBM was where Mr. Buffett wants his money to be. Maybe he knows something I don't. He has more access to Ginny than I did. And he certainly has been able to find lots of companies that aren't exactly cutting edge (for example BNSF and GEICO) that have a great long term story. But, personally, I just don't see IBM being able to preserve the current levels of profit and margin long-term.