Future Horizons industry midterm forecast seminar20 July 2010, Kensington, London

Citation

Ling, J. (2010), "Future Horizons industry midterm forecast seminar20 July 2010, Kensington, London", Circuit World, Vol. 36 No. 4. https://doi.org/10.1108/cw.2010.21736dac.005

Future Horizons industry midterm forecast seminar20 July 2010, Kensington, London

Article Type: Exhibitions and conferences From: Circuit World, Volume 36, Issue 4

Future Horizons have been in operation for 21 years, so congratulations were in order. Malcolm Penn, CEO, kicked off with a look at the economies. In Q1 this year, the world GDP expanded by 4.5 per cent per quarter, and for 2011 we will see 4.25 per cent per quarter. Advanced economies are modest, and unsteady, whereas developing economies and much stronger and steadier. A double-dip recession? Not sure about that, but the European debt crisis will slow the pace of global recovery; confidence is low, and the downside risks have increased.

Global recovery requires strong policies, and inflation is to remain subdued. Living within ones means is the new fashion, this is obligatory as there is now no cheap and easy money available, but this does not apply to emerging economies where internal demand will help sustain recovery.

In 2010, World output will be $62.7 trillion. The USA, China and India have bounced back the hardest. Malcolm thought that we had had a disruption, not a recession. Why? He explained this later on. Back in the 1990s, there was a divergence, with emerging economies growing much faster than advanced economies, in fact twice as fast. Interest rates will remain low for the next six months, but if left uncontrolled they could trigger the next bubble; it has to go up, and the IMF think that it will be another 12 months before it does.

The forecast for 2011 is 3.0-3.5 per cent GDP growth, as good as it was back in 2004. This applies to the USA, less so in Japan (2.4 per cent) where public debt is huge, and the yen is still high. As are the unemployment figures, but the potential for change is enormous.

China and India – here the figures are rather different, as you might expect. China’s GDP is growing at 10.5 per cent this year, India at 9.4 per cent. China is troubled by high inflation and growing labour unrest, and India is still suffering from chronic infrastructure problems, and inflation is another headache. Back in troubled Europe, we might struggle to get 1 per cent GDP growth this year, and that is the lowest growth of any advanced economy. EU unemployment now stands at 9.7 per cent (23.3 million people).

In terms of world economic outlook, it is confidence rather than optimism that is gaining momentum, building on the foundations of recovery; recessions (or disruptions) can lead to innovation, which in turn attracts investment.

Industry outlook

The Four Horsemen of the Apocalypse are still in harness, and they pull the Future Horizons visionary chariot with the same accuracy:

1. 1.

   The economy. The FH research philosophy is that you can model all you like, but there is no substitute for experience and intellect, and knowledge of how the industry works. Semiconductor industry forecasting is not an exact science and never ever think that it is, but understanding the real growth is a fact of life, and now we have 34 per cent growth. Or do we? What we have had is not a normal boom or bust cycle; the numbers we are seeing are purely mathematical and have nothing to do with the market. It was a motorway crash, a disruption, which lasted for three quarters. The reality is that there has been continued growth and the chip industry has been back on track since Q3 2009. Malcolm cautioned against looking at numbers by themselves, they are possibly meaningless; one needs to look behind the numbers, as the meaning can be radically different. He described how the comparison between Q3 2008 and Q3 2010 demonstrates how the emerging market drives the recovering world economy; the strong unit demand with no excess inventory; the tight wafer fab capacity and better still average selling prices (ASPs) are recovering, structurally driven. The differences are in the detail, not the main headings.What we are now looking at is 36 per cent growth, with the market due to break though the $300 billion threshold in December, this year. In a free market economy, things happen that are not logical, such as reducing ASPs for products that are in short supply.

2. 2.

   The demand. Is unit demand predictable? Monthly – no. But the long-term trend IS, it is 11 per cent per annum. The chip market is in fact a collection of three big markets, with a lot of $5 billion niches. It is an immature market, and as such it is still growing.

3. 3.

   The capacity. When it comes to manufacture, it takes a very long time indeed to increase capacity; as a minimum it is one to two years to bring a new line into production. The business model for fab capacity is in fact broken, it is now a “stop-start” situation, and the industry is doing absolutely nothing about it.

4. 4.

   The ASPs. The least understood industry wild card. ASPs are driven by new designs, but that takes two to four years to work through. If excess capacity kept prices low, this is no longer the case, and the trend is now upwards, the right side of 10 per cent.

Key market drivers – there is a five- to ten-year gap for products to come to maturity, and so for 2011, FH have said that growth will be +14 per cent as the rebound was in fact much faster than had been expected, and they started up again from where they left off. There will be an overbuild, we will see it coming, 2011 will see the peak of the structural cyclical boom, and by 2012 the market correction will be in full flow, so by 2013 there be the next cycle of overcapacity. From 2010 to 2012 will be good years, so make the most of them.

A blend of industry and government issues might be healthy when it comes to the location of manufacturing capacity; security of wafer supply is crucial, and the changes that have occurred in the world with the location of chip manufacturing factories means that unless something is done about it there may be just three major chip firms for the future high-technology semiconductor needs. That is unhealthy. Not getting your chips when you want them is a real threat to the survival of your company, needless to say, but the industry still seems to go from feast to famine with worrying regularity. We are spending no more on R&D than we were 20 years ago. Why? Too many chip firms lack global aspirations, they are looking for the “quick fix” and the “short route to money” – risk aversion is not risk management. The supply chain system needs to talk to each other, if disastrous shortages are to be avoided. Supply chains are not infallible.

Key application markets

Mike Bryant is the CTO of Future Horizons, and he looked at the ways the market is going, which applications are what, and where are the emerging opportunities.

Growth in electronic equipment is both from developed and developing markets. Demand outstrips supply, with the automotive sector still down but recovering well and as usual PCs and mobile phones lead the market. The year of the netbook was 2009, and business renewals running late but now picking up. Government IT contracts will be on hold for more than a year. Laptop battery life 64 bit apps may overtake 32 bit apps. But they will need low-power servers; the power consumption of these (Google, Amazon Microsoft, etc.) is truly massive, and reducing the costs of powering and cooling these servers quickly pays for a lower power update. Of 40/100 Gb Ethernet links will replace 109 Gb in many places.

Three-dimensional monitors in the home is a “must-have” for the future. IPv6 will be needed soon, but with the rise of smart phones means we are really about to run out of IPv4 addresses. There will be a growth in networking replacement equipment. By 2015, 300 million laptops per annum will be produced.

Mobile phones – smart phones are a “must-have” fashion accessory, and mobile phones are have shown excellent resilience to the economic downturn. There is still relentless pressure on prices for semiconductors, and amongst the suppliers of these MediaTek., Qualcomm, Broadcom and ST-Ericsson will survive as the main suppliers, with a second tier of Marvell, Infineon and Ti. The iPhone 4 launch is a disaster; the direction Nokia is taking seems uncertain, and whilst Blackberry continue to have excellent products they are still losing ground to Apple mobile networks can no longer cope with the volume of data demands from the smart phone use, and charging may change to a “per gigabyte” basis to slow demand.

Future trends? The “Always On” GPS is one, with more powerful graphics and powerful application processors in smart phones; wireless link speed is still the limitation and will possibly never be resolved. Maybe, Apple or Google could push this. Mobile phone unit production shows no signs of diminishing; 1,250 million units by 2014.

Other hot spots include saving on power consumption; consumer markets as ever dominate; the industrial market will offer opportunities including bio-fuel convergence of electronics and biotechnology. Security for home environments; medical applications such as the “lab-on-a-chip” concept; remote medical analysis and monitoring; the age of mobile data may finally come to pass with the convergence of mobile phones and computing, and landline-based services are also due for a major upgrade.

Main markets – 3D TV may be an upcoming market driver. Standard LCD TV sales remain strong. LED TVs are taking an increasing market share at premium prices, with 2010 sales expected to be around 180 million, rising to 228 million by 2013.

The arrival of the iPad was one of the major events of the year, but competitors are now seeking to launch their own versions. Bluetooth 4 will allow very low-powered Bluetooth wireless products and sensors using coin cells. Near Field Communication – can be used for payments, security and other contact less communication, and will eventually be in every mobile phone.

Fast internet – apart from a few far-sighted countries such as Korea who invested in fibre networks, current broadband networks are failing to meet user expectations, now many governments are considering mandating minimum standards in these networks.

In terms of semiconductor technology, Mike was at pains to point out that Moore’s Law is not coming to an end. In fact, there is now a valid addendum – the cost of development of a new process is roughly inversely proportional to the node length. So the cost of developing 22 nm technology was twice that of 45 nm which was itself twice that of 90 nm etcetera etcetera. Developing new processes is becoming more expensive, and developing new designs using that process is becoming crippling so. Yet, the price per square centimetre of a silicon die remains constant.

The effects of Quantum Mechanics dominating Newtonian physics were changing the game, once we get to the 10 nm chip, and finding engineers comfortable with quantum mechanics might be difficult. However, there are alternatives to silicon-based CMOS, here strontium-germanium inter layer’s between a hi-k insulator and a germanium channel can give us several more generations, but beyond that we may have to go to carbon-based transistor on silicone, or a benzene ring. SOC designs will have to be fully fault tolerant, or reconfigurable, so that faults can be reworked, but the number of processing steps will be the practical limit, as there will be very many of them.

Future Horizons always have well-attended seminars, and when you have been to one you can see why. In the USA, they have a phrase about bangs for bucks, but in the UK I suppose we might prefer to call it extremely good value for money, and it is all about as close to a gaze into a crystal ball as we are likely to get.

Notable future events include the next Industry Forecast Seminar on 14 December 2010, again at the London Hilton Hotel, Kensington, and the International Electronics Forum which is to be held on 9 May 2011, venue TBA but likely to be most attractive, we are informed!

John LingAssociate Editor