TSMC , Smartphones and the Internet of Things: Behind the Numbers

              

On March 27, TSMC’s venerable chairman Morris Chang spoke at the annual Taiwan Semiconductor Industry Association meeting in Hsinchu. In his speech, Chang delivered a number of interesting announcements and prognostications, including:

1      - Cellphone sales would grow 4.9% year to year from 1.8B in 2013 to approximately 1.9B in 2014;

2     - Worldwide smartphone shipments would grow 25% from 2013’s 968M to 1.25B in 2014;

3     -  Smartphones would continue to be the leading semiconductor growth market for the next two years;

4     -  The 28nm and 20nm nodes would be the primary beneficiaries of that growth;

5     -  Moore’s law has another 5-6 years of life left in it;

6     - The Internet of Things (specifically, mobile computing devices other than tablets and smartphones) would be the driver of semiconductor unit and revenue growth after Smartphones, lasting for 5-10 years

7     - The most important semiconductor technologies and skillsets for the IoT era would be advanced chip packaging services, MEMs sensors and low power technologies & design methodologies that could reduce power consumption by at least an order of magnitude.

Concurrently, communiqués were issued by AT&T, Cisco, GE, IBM and Intel regarding their partnership in the newly formed Industrial Internet Consortium. The group has tasked itself with developing specifications and standards for interoperability, testing, modularization and security of IoT products targeting applications across a broad swath of industrial markets.

One can infer a certain synergy and theme of optimism in these back to back announcements. Yet before acting on these news items, a sober analysis of the facts and assessment of risks is in order. This editorial will constitute that critique, separated into two parts.

Part 1: The Mobile Phone Market

A critical analysis of Chang’s predictions can be performed by taking a closer look at the numbers. Below is a table of mobile phone unit shipments over the last six years. The source is Gartner, the same provider of general market data that TSMC uses.

		2008	2009	2010	2011	2012	2013
smartphone units (MM)		128	172	297	477	680	968
Growth (y/y)			34.4%	72.7%	60.6%	42.6%	42.4%
Feature phone units (MM)		1094	1039	1300	1299	1066	839
Growth (y/y)			-5.0%	25.1%	-0.1%	-17.9%	-21.3%
mobile phones (MM)		1222	1211	1597	1776	1746	1807
feature phones (MM)		1094	1039	1300	1299	1066	839
smartphone units (MM)		128	172	297	477	680	968

One can readily discern the healthy growth of smartphones and the concurrent decline of feature phones. Yet the overall mobile phone market has been essentially flat for the last three years. The growth of smartphones is clearly stemming from the direct cannibalization of feature phone accounts. This effort went into high gear during 2013, as smartphone unit prices declined in what many observers deemed a price war and service plan costs shrank in parallel to entice cash-strapped consumers to switch to a higher functionality mobile phone.

The TSMC forecast for 2014, however, raises some eyebrows. Chang predicts growth of the entire market by 100M units for a total of 1.9B mobile phones and a further 25% growth for smartphones to 1.25B. Simple subtraction tells us that Chang is expecting roughly 675M feature phones to be bought in 2014. Doing a little math, that means the feature phone market would decline by 19.5%.

Now we can see contradictions emerging within the detail of Chang’s forecast. Smartphone growth is plainly rolling over, as the historical numbers suggest. Even the 42.4% growth rate of 2013 belies a growing weakness in the sector, as the second half of the year saw smartphone unit and service plan pricing promotions reach a panic pitch in an effort to maintain flagging momentum. Market observers are seeing parallels to the PC market as it matured in the 90’s, with features and enhancements losing the ability to command value, forcing manufacturers more and more to compete on the simple basis of price.

Since smartphone sales are evidently growing thru cannibalization of feature phones and feature phone sales are historically declining at an INCREASING rate, one would expect smartphone sales growth to correspondingly weaken and the overall market to either stay flat or decline. Yet Chang’s prediction calls for feature phone momentum to REVERSE and the overall market to GROW.

Why? There’s no apparent logic to it if one looks at market conditions. The world economy in 2014 is not strengthening over 2013, but manifestly declining. China’s economy has reversed momentum and is rather clearly headed for a recession – possibly a strong one. Exports shrank by 18% last year – a first for China in many years. All the BRICS nations are suffering as well - their exports have declined precipitously and growth rates have tapered, stalled or even reversed. Copper and iron prices worldwide are deteriorating to multi-year lows – highly indicative of a major slowdown in worldwide manufacturing. The western economies continue to stink to high heaven as they wallow in a financial cesspool of consumer, financial sector and sovereign debt, and the previous illusion of immunity projected by the NYSE’s record year belie the fundamental weakness of the US economy – a fact which the deteriorating housing market and deeply depressed workforce participation rate (stuck in a 35 year low) are revealing for all to see. America is evidently following Europe and Japan into economic stagnation and fiscal deterioration.

All of the above indicate that consumer discretionary income is not going to grow in 2014, but more than likely shrink even further. This suggests that TSMC’s expectations are unrealistically optimistic both in terms of smartphone growth and feature phone decline. A more realistic forecast would be one in which the mobile phone market would be lucky to stay flat against 2013’s numbers and more likely than not to experience a small decline.

Part 2: The Internet of Things (IoT)

Morris Chang expressed high optimism that the latest industry efforts to create portable computing devices beyond tablets and smartphones would lead to a wave of highly attractive consumer products that would continue the revolution in how people across the world live, work and interact. It would be rather stupid to discount Chang’s views out of hand, as he has proven to be one of the two or three most prescient and insightful leaders in High Technology over the last thirty years.

Nevertheless, the formation of the Industrial Internet Consortium (IIC) should not lead one to conclude that the industrial markets targeted by IoT device developers will shortly burgeon into an electronic renaissance with stellar growth and precipitously ramping volume. That the formation of broad industry partnerships to standardize electronics markets is often beneficial cannot be rationally disputed – just look at the long term effects of the IEEE’s 802.3 and 802.11 specifications on wired and wireless connectivity. Yet the IIC’s area of interest is so fragmented in terms of covered markets, technology and applications that it cannot help but give one pause.

Consider the markets which the IIC has chosen to encompass with a set of unified standards, specifications, testing guidelines, modularity and security protocols: energy, medical, manufacturing, transportation and the public sector. Each of those markets has multiple subsegments with their own application requirements and end user markets. The overwhelming majority of opportunities are low volume sells that require high levels of software and hardware customization. There are also a great many trade secrets wrapped up in medical equipment design, manufacturing processes and energy exploration – secrets which sector participants often think are integral to their value proposition. Such companies will be inherently resistant to revealing even a hint of what their ‘secret sauce’ might be like, further hobbling the efforts of standards definers.

IIC members seem to be distinctly aware of the obstacles they face in making their organization effective. When queried as to what specific initiatives there were on which the group would begin work and when the first set of standards could be anticipated to be released, the general response was “Ummmm……..errrrrr…….yeah, uuuuh, we’ll get back to you on that.”

Does this mean that we won’t be seeing some sort of revolutionary IoT device or set of products hit the market two years from now, as Morris Chang implied? Well, anything is possible. Genius is not a regularly occurring and predictable phenomenon. Nevertheless, it’s abundantly clear that no such products exist today. With this in mind, it might be a safer course of action to assume that Morris Chang was more than anything simply expressing his hopes rather than issuing a forecast on the IoT and its influence on the medium and long term prospects of the electronics sector.