Tom Armistead

About the Author Tom Armistead

I'm a well-informed retail investor and post on SA in order to expose my thought process to critical examination and comment from readers. It makes me a better investor. I'm particularly proud of bullish macro articles posted in 2009 and later, in which I presented ideas that encouraged me to invest very profitably in a rising market. I also did articles on individual stocks, many of which contained insights not available elsewhere. Finally, I wrote a number of thoughtful articles critical of financialism and the lack of ethics on Wall Street. I do not post for compensation, as I am concerned that editorial policy encourages and pays a premium for articles that invite the reader to speculate on the short term movements of microcaps, penny stocks, and controversial issues. The best way for me to monetize my insights is to invest accordingly. As a retail investor, I don't give investment advice. I write about what I'm investing in, and the thought process involved in decision making and stock selection. Hopefully some of what I write is of benefit to others, by sharing my experience as I interpret it and helping them improve their investment thinking and process.

IBM: Taking A Hard Look At R&D

International Business Machines (NYSE:IBM) is cheap by many of the usual metrics, thanks to ongoing lack of revenue growth. The long-term potential created by the company’s ongoing investment in R&D is not being recognized in the market. This article takes a hard look at IBM’s R&D spending, and concludes that it is appropriately targeted, and more likely than not to lead to renewed growth and continued leadership in enterprise computing.

The stock has been pummeled following the announcement of the sale of the microelectronics business to GlobalFoundries, as well as the poor 3Q 2014 and withdrawal of long-standing guidance for 2015. Before going over the R&D issue, it makes sense to deal with the most recent divestiture.

Sale of Microelectronics

According to CFO Martin Schroeter on 3Q 2014 the earnings conference call, the chip business didn’t have enough scale to compete with the likes of Intel (NASDAQ:INTC), Taiwan Semiconductor (NYSE:TSM) and GlobalFoundries. As the industry moves to progressively smaller nodes, the investment required in capex and R&D increases, and the demands on capital became excessive compared to the funds available from future business. I was able to verify the lack of scale by consulting the segment information in IBM’s 10-K.

IBM used its foundry operation for its own chips and needed to ensure that the operation would continue so that the output would be available for servers and mainframes. While the R&D staff supporting the operation were a prime motivation for GlobalFoundries to do the deal, IBM will continue to enjoy the fruits of their labor under a 10-year agreement that gives them access to product at the 20 and 10 nm nodes. Hence the payment to the buyer.

IBM will continue its $3 billion over a five-year commitment to 7 nanometers and beyond, together with other projects designed to deal with the difficulty of extending Moore’s Law at 7 nanometers and below. The company believes newtechnology will be required, and they’re working on it.


The annual expense runs about $6 billion, or $6 per share. Capex, net of depreciation, runs about $1 per share, so the bulk of the capital that the company allocates to grow the business is deployed into R&D. The question comes up, what does the investor get for this large and ongoing expense?

John E. Kelly III, IBM’s Director of Research, along with Steve Hamm, a writer for the company, has written a short book entitled “Smart Machines – IBM’s Watson and the Era of Cognitive Computing.” As a way of getting an understanding of the company’s R&D focus, I bought the book and read it.

A New Era?

Chapter 1 makes the case for a new era in computing, and specifically an Era of Cognitive Computing. In the future era envisioned, computers will more closely approximate the human thought process, with the advantage of the ability to access vast amounts of data and computational power. These machines will interact as partners with humans, vastly improving the accuracy of medical diagnosis, customer service and financial analysis, as examples.

This vision runs up against the technological limitations of current architecture, as well as limits on how small microprocessors can be made. While substantial progress can be made by incremental improvements, at some point something totally new will have to be created. If there’s a needle in the haystack here, IBM will find it.

At the same time, R&D also is directed at what’s happening in the here and now – in cloud, social, mobile and security. They view these opportunities from their vantage point as a leader in enterprise IT.

Big Data and Analytics

Huge amounts of data, much of it in unstructured form, are created daily. Extracting insights from this complexity is a daunting task, and a sizable opportunity. As an example, human analysts peruse 10-Q’s, 10-K’s, earnings call transcripts, presentations, and industry specific literature in an effort to identify above or below average investment prospects. Connecting the dots isn’t easy, but the rewards can be substantial. Watson can learn to do it: anything much above 50/50 is huge.

The following slide from a June 2014 IBM presentation is instructive:

Screen Shot 2014-10-22 at 18.42.09

The point here is that the current hot spots of mobile, social, cloud and analytics cluster around the central concept: big data. That’s IBM’s view. And they have been all over all of it.

Relevant R&D

Assessing the connection between R&D and commercial implementation is difficult. 3M (NYSE:MMM) does a good job making the information available in a form that is useful to investors: however, as an industrial business it’s easier to quantify.

IBM addresses the issue in several recent presentations: Two-thirds of R&D is deployed into data, analytics and cognitive computing. They have 1,500 + cloud patents, as well as 4,300 in mobile, social and security.

My take: IBM is spending money where it will improve future earnings. There is no reason to label the R&D spend as a waste of time on ivory tower academic make-work.

Another thought: IBM has fabulous ROE and ROIC metrics. It is a stretch to believe that a company that has a history of deploying capital into profitable areas will not realize fine returns from its R&D investments.


Before discussing the area of mainframe, it will be useful to see how much it is contributing to the company’s earnings and revenue. Briefly, the STG (Systems and Technology Group, which includes mainframe) segment lost $503 million for 2013, and contributed 14% to revenue. This segment, which also includes servers, is the source of much of the company’s revenue decline, and is not pulling its weight in contributing to overall results.

The point is, both servers and mainframes are entering a refresh cycle. Having sold off the commodity X86 business, IBM has some fine offerings in both the Power8 server family and in mainframes.

Over time, I look for this segment to contribute to both revenue and EPS growth, starting from a low base, following the divestiture of the X86 and Microelectronics businesses.

The Big Cycle

My father had a successful career in R&D, and in his later years enjoyed telling war stories, some of which contained usable insights.

Looking at the history of Corning (NYSE:GLW), where he worked, he saw a cycle of innovation, profitable growth, commoditization and obsolescence that ran slowly and inexorably over long time frames. One example: light bulbs. When Edison got his bright idea, he needed someone to bring it to commercial reality, and make the bulbs to sell for $1. That was the Corning Glass Works. It was a wonderful business, but eventually became a commodity, and the product is now obsolete. Corning divested the operation at the commodity stage.

Another: cathode ray tubes for TV. Corning figured out how to make them, and made wonderful profits as the age of TV developed. Again, they exited as the commodity stage gave way to obsolescence.

IBM famously created the Personal Computer business, which flourished in part due to the open architecture feature, which encouraged an innovative industry, not to mention competition from clones. They left the business when it became commoditized. Of course Intel is still making very good money on PC chips.

The cycle of innovation, profitable growth, commoditization and obsolescence doesn’t run nearly as fast as some pundits think. IBM plays at the front end of this wave, and sells the commoditized and obsolescent businesses without regret. These next big thing for businesses does not come along that frequently, and the innovative business model consequently produces lumpy results, over long time frames.

I don’t think the cloud, mobile, social media, or the internet-of-things define the new era as cogently as the concept of Cognitive Computing. Maybe it’s really an era of Big Data. In any event, winning a naming contest doesn’t imply commercial dominance.

It’s early in the new era, and we are a long way from knowing who the winners will be.

IBM is about Innovation in Enterprise IT

Getting back to the idea that IBM is a cash cow/buyback machine, that’s how the shares are priced right now, and very conservative even on that basis. From the 2Q 2012 10-Q, here is how the company describes itself and its strategic actions:

The underlying theme of these actions – from the expansion of the cloud platforms and capacity, to the OpenPOWER consortium, to the partnership with Apple for enterprise mobility, to next generation chip technologies – is that the company is leveraging its unique strengths in innovation and enterprise capabilities to maintain its differentiation in the emerging areas of enterprise IT. While some of these actions impact results in the short term, they position the business better for the long term.

The point is, the investor gets the whole R&D operation as a bonus. Taking a long term view, at some point in the future innovations are going to create profitable business opportunities that will ignite growth in revenue, with a corresponding increase in earnings.

In the meantime, patient investors can collect the dividend, and watch the buybacks create value.

Valuation, Risk and Reward

My preferred PE5 method suggests a target price of $215 by the end of 2015. Simple models based on growth in the company’s strategic imperatives – analytics, mobile, cloud, social and security – with a slow decline in other areas demonstrate the possibility that the company will return to overall top line growth over the next several years. That will take place from a lower base, if it does occur.

Mediocre growth, or no growth, can still generate a rising share price, provided buybacks continue to create rising EPS. This works better when share prices are low, as they are now. Projecting GAAP EPS at $16.50, and applying a P/E of 11.7, my low case is $193, within two years, and I’m holding on that basis.

If we do get a new era in computing, and IBM is properly positioned with what it takes to get down under 7 nanometers, and support Cognitive Computing as envisioned in the book, the payout would be much larger, and can’t be estimated with any credibility. I want to have a seat to see how that plays out.

Caveats and Reservations

It is entirely possible that a disproportionate amount has been spent on buybacks, at the expense of R&D and capex. It can be argued that the company starves good businesses to death and then sells the residue.

Why return capital to shareholders if it can be deployed into growing the business? CEO Ginni Romnetty has been relentlessly executing according to plan, and seems baffled that results haven’t been as projected.

Hewlett-Packard (NYSE:HPQ) seems to have done better since Ralph Whitworth of Relational Investors was on the board. It’s too bad he had to leave for health reasons. It would be nice if someone of his stature could join the board to provide some fresh outside perspective. Meg Whitman might have some insights.

I’m long the stock, and moderately optimistic. However, I would be more so if the pace of R&D and capex were increased, and buybacks held to some lower amount. Recent buybacks were funded in part by an increase in non-financial debt, a slippery slope.

According to, which measures analysts’ and bloggers’ success rate based on how their calls perform, analyst Tom Armistead has a total average return of 14.2% and a 68.4% success rate. He is ranked #309 out of 3906 bloggers.

Stay Ahead of Everyone Else

Get The Latest Stock News Alerts