IBM thinks again

IBM had an annus horribilis in 2014, but has been incubating its future in artificial intelligence (AI). Big Blue’s stock fell almost 16% between October 1 and December 16 2014 as it missed earnings forecasts and reported a 10th consecutive quarter of falling revenues.  The $5 billion expansion of its $13.5 billion buyback programme enraged some analysts, and the jettisoning of the chip business to Globalfoundaries was met with a mixture of relief and resignation.

To put this in context, IBM remains the world's fifth most valuable brand and is a $160 billion company with a huge patent portfolio which generates more than $1 billion a year in pure licensing revenues. Warren Buffet has IBM as his third largest holding because of these factors. Then again, in 1999 Nokia’s market cap was $245 billion, it was a top-five global brand and boasted more than 10,000 mobile patents. As Andy Grove might say, given these facts it is okay to be a little bit paranoid.

Unlike Kodak or Blackberry, IBM is not suffering from complacency, but quite the reverse: it has been in revolutionary paranoid cost-cutting mode since Lou Gerstner took over in 1993. But the mantra of teaching elephants to dance, shedding hardware businesses and buying back stock is of limited value when your core markets are dramatically changing, which is the reality undermining IBM’s future.

Recognising this, IBM has made a big push into the three areas where its future lies: services, the Cloud and software. The acquisition of PWC Consulting in 2002 built a critical mass of know-how and brand, but the business is still poorly integrated and crucially the model for success has changed. Clayton Christianson has commented how consulting is being disrupted (https://hbr.org/2013/10/consulting-on-the-cusp-of-disruption) as automation and AI replace offshoring and Moore’s Law makes fixed-price IT contracts start to look expensive for clients. The Cloud’s biggest impact is on the huge small and medium-sized segment, which traditionally IBM serves poorly. The value for users is lower costs through software as a service, but IBM seems to be counting on high margins without proprietary software to deliver to this segment.

One business strategy for futureproofing companies is to make asymmetric bets on new technologies or business models and then acquire the successful offspring. Corporate venturing is the fastest-growing (and best-performing) venture captial strategy for this reason.  Google is perhaps the most prominent exponent of this strategy, but more than 80% of companies listed in the Dow Jones Index have established corporate venture units and IBM is no exception – incubating a significant asymmetric bet in the shape of Watson. 

Watson is known to many people in a trivial sense as the machine that won the US game show Jeopardy in 2011. However, after almost a decade of work, $1 billion of new funding in 2014 and a $100 million venture fund to catalyse an ecosystem around Watson, it is clear that IBM’s commitment to AI is highly strategic. The market for cognitive computing is set for exponential growth and Watson is at the forefront of this. Deloitte forecasts a $50 billion market in the next three years alone, with global applications doubling this figure.  The hidden value in Watson and its intellectual property has the potential to be extraordinary.

I had the pleasure of hearing Watson present in December 2014 at Sanford Bernstein’s Tech Innovation Conference in Boston. Its roll-out into healthcare research and development and financial services, in particular, caught my attention. Applications where vast amounts of data need to be trawled to find customised solutions for clients are a sweet spot. Initially deployed as an adjunct for human beings, it is not so difficult to imagine Watson replacing analysts or financial advisers in the near future. Recently both Steven Hawking and Elon Musk have spoken out about the dangers of AI making people obsolete, and both are at the forefront of AI projects.

Irrespective of the future impact of AI, what is clear is that the pace of change in AI-related innovation is dramatic. More than 12,000 patent filings directly reference AI and the rate of growth is increasing. There are innovations in almost every field, including components like memsistors, batteries, robotics and big data. Venture capitalists are active in this sector, with more than 170 new start-ups in AI being funded from Silicon Valley. The AI value chain is also migrating relentlessly towards neuroscience-based deep learning, because it can adapt to new information and learn in the same way a human brain. This is an area where IBM, Google, Facebook and a handful of others have a headstart. 

Thomas Frey, head futurist at Google, recently observed that in the 1960s the average holding period for a stock was four years. By 2000 it was eight months and by 2008 it was two months. Today it is 20 seconds. The reality is that AI is already deployed in a meaningful and profitable manner in the capital markets, initially through a handful of hedge funds such as DE Shaw and now on a systemic scale and with good reason. In 2014 AI-based trading strategies were the most successful alpha generators of all with Lord Stanley Fink’s ISAM up more than 62% for the year, more than four times the return of the S&P 500. Interestingly, DE Shaw’s founder David Shaw now dedicates himself full time to applying AI to medical research, focusing on computational biochemistry at DE Shaw Research and Columbia University, so cross-pollination is active.

So where does this leave IBM? Under Gerstner, Palmisano and now Rometty, IBM’s relentless focus on shareholder value has alienated many staffers. Former employees often mention an erosion of culture and loss of know-how as talent leaves the organisation and internal skunk works are closed. Now that the holy grail of the $20 EPS target has been abandoned, the cash cow that continues to provide Buffet with an 'equity bond' may find that ironically its most valuable investment has been one of the few remaining skunk works in-house, a group of game-playing geeks in the basement at Watson.