Before describing IBM's opportunities, I will look at the potential downside which should always be of a primary concern for an investor (I have elaborated more on this in an article). I will use information from a recent transcript in various places.
Folks like Alsin and Druckenmiller are predicting doom and gloom for IBM because of possible margin declines due to increasing adoption of cloud computing. In addition, many analysts are focused on IBM's declining revenue. Warren Buffett stands on the other side: Berkshire Hathaway (BRK.A, BRK.B) invested about $10 billions in IBM at prices similar to those available today.
Let's start with the cloud threat. First, IBM is a serious cloud player itself, one comparable with Amazon (NASDAQ:AMZN). IBM's cloud services are already cheaper than AWS---and if Amazon's business model in cloud is similar to that in retail, then AWS results in more losses than profits and will have to increase prices in the future, losing competitive strength.
The point of the IBM's interest in cloud is not in competing in low-margin IaaS, but in the integration of a public cloud with private records of companies, exploiting their client relationships and access to private records which cannot be easily replicated by competitors. One third of IBM's R&D budget goes into this, and I believe IBM has a competitive advantage here. Their business is rather sticky (it is complicated to change your IT provider) and will be even more sticky after a successful integration of this kind. In addition, IBM offers much more than a generic cloud infrastructure provider can offer. Many businesses need specialized non-trivial applications; I doubt very much that companies like Amazon have experts ready to tailor their cloud applications to suit particular client's needs.
Second, IBM's revenues have missed analysts expectations in six quarters in a row. Apart from one-time issues like currency translation and some political changes in China, this was intentional by IBM: they decided to exit unpromising low-margin businesses and focus on increasing earnings instead. I am pleased with such a decision as an owner of the company. The management is wise enough to keep a reasonable amount of hardware business to complement their software and services offerings.
It is important to understand that while IBM could use their cash to buy revenue in the form of established larger businesses, they do something rather different. IBM is engaged in many small acquisitions, buying intellectual property to fill the gaps in their portfolio. They are very disciplined in this, and aim to buy for the future. Such an approach is rare among companies and underappreciated by Wall Street.
Financially, the company is doing OK, generating rising amounts of free cash flow (check an article by Early Retiree if you are interested in numbers). There is very little debt apart from their financial arm which invests in receivables. The analysts are concerned that earnings will stop to grow because revenues stay flat. Even if that happened, an investor would get about 9% free cash flow yield in the form of dividends and share buybacks, which I view as the worst case. (Yes, the share price can decline to something like $120 temporarily, but all the better with more than $20 billion allocated to repurchases.)
One third of the R&D budget goes into analytics. There are huge amounts of data collected by enterprises and governments, and there is underlying value in the data. While I do not think IBM has a particular competitive advantage here compared to, say, Google, the market for big data analytics will grow in the future and IBM will keep a reasonable share of it. I believe it will more than offset declining revenues from IBM's legacy businesses.
The final third of the R&D budget goes into Watson. And Watson is what gets me really excited about IBM. It is better to keep emotions out of investing, so let's try to keep rational.
Centuries ago, there were people called polymaths, capable of becoming leading experts in almost all science areas. Nowadays, it is impossible; human capabilities are genuinely limited. For instance, medical research produces more studies than doctors could follow even if they did nothing except reading. As an investor, you have surely experienced that following just the few companies in your portfolio easily takes all your allocated time. Note how quickly new information is priced into stocks: it usually takes only a few trades for most liquid tickers. However, if a human is to act quickly, it often leads to a compromise on quality; investors are known for overreacting to news. Imagine you can hire a perfectly rational super-fast learner that would aid you in your decisions, one that can read websites, articles, annual reports; you can teach him over short period of time in what way to process information and how to present it to you. You can even let him make the decision for you. And this is what Watson aims to be.
Several studies have shown that simple statistical models often beat experts in the area of their expertise. This can be particularly true for doctors. A wrong diagnosis may cost quite a lot, and so it pays off to pay a lot for a good advice. As a doctor, you don't have the time to look at millions of patient records or follow the latest research on various drugs. And usually you don't have money to pay for a large team of qualified people doing it for you. Consequently, I believe there are ample opportunities for systems like Watson, capable of reading unstructured materials and learning from them. The future is unclear---maybe doctors will keep the system to themselves, or maybe nurses will act on diagnoses provided by Watson---but in any case, Watson is likely to become an important element of the healthcare chain.
I especially like Watson from the moat perspective. If you have a system already trained in oncology, you can offer attractive prices for services to additional oncologists. It is likely that an artificial intelligence system will produce blunders from time to time, but with more feedback the system will improve and mistakes will be less and less common. The network effect could be particularly strong: as a doctor, you put your reputation at risk, and so it is less likely you will use some start-up system with unknown reliability. The patients' demand will make you join the train even if you resist; it is hard to answer questions about why you do not use top technology as your peers do. The first mover will gain a huge competitive advantage here.
Notice also that Watson's potential revenues would come from budgets not used for IT services before.
I perceive IBM as a great long-term investment. Their cloud integration efforts and big data analytics combined with present software and services offerings should lead to stable revenues and profits. In the worst case, one gets 9% free cash flow yield with little or no growth. I consider it realistic to assume 2% inflationary growth here, leading to an 11% total return.
In the best case, Watson will really take off in various industries in a decade or so, leading to multiples of current profits. I think it would be just dreaming to try to estimate the upside precisely, but I have a strong faith that Watson will become a part of several medical procedures in 5 to 10 years, leading to a few additional billions of profits. I wouldn't be surprised with 12 to 15% long-term returns.
IBM truly is a business of the type Warren Buffett likes: throwing off plenty of cash, having little potential downside, and diligent at widening its moat.
Disclosure: I am long IBM, BRK.B. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.