If only wishing made it so…
By Luke Collins

29 April 2008

The chip business is tough. The technology is difficult, the capital costs are horrendous and demand is so uneven that it makes a ride on a fairground waltzer seem like a quiet night in. Little wonder then that senior execs are driven to hoping that if they just wish hard enough for something, it’ll come true. It must sometimes seem as valid a strategy as any other.

Latest victim of the wishing disease is Moshe Gavrielov, the newly installed CEO of programmable logic maker Xilinx. He faces a couple of issues. The first is that the market for programmable logic is limited. The chips tend to be big, hot and expensive, which limits their uses and Gavrielov’s ability to promise faster growth.

The second is that he’s a new CEO, so he needs to look busy. His plan, then, is to expand Xilinx’s business through a ‘Goldilocks strategy’ of defining application-specific products that are ‘just right’ for various markets.

He told a recent conference that the total available market for programmable logic will be worth $5.2bn by 2011, but that the market opportunity for parts that include digital signal processing or embedded microcontrollers, or which are highly targeted application-specific standard products, would be $14bn.

“My expectation is that as we define our products more carefully we’ll be able to address more of that market,” he said.

The problem is that the FPGA community has been trying to make chips that have an ideal combination of programmable logic and application-specific elements, such as SERDES blocks for the communications markets, for years. But these parts often have exactly the wrong combination of features for many applications, being too expensive for simple projects or lacking vital features for more complex projects – the Goldilocks problem.

“When you come from an approach that is technology-driven and move to a more market-driven approach then you are more likely to fall into those traps,” said Gavrielov. “Making the right product definitions will require a very delicate approach, and being close to our key customers. Is it simple? No. Will there be cases where we overshoot by 10 per cent? Yes, but that is less of an issue.”

Well, maybe. But even if Xilinx can define a series of chips for market niches which are “just right” – not too large to draw the attention of the ASSP vendors nor too small to be worth going for – there’s still the stockholding problem. The great joy of making programmable logic is that the chips are high margin and apply to many markets. If the communications industry is tanking, call the computer makers. Once you start making application-specific programmable logic, you’re locked into the fortunes of market niches, stocking wafers and making long-term promises that are expensive to keep.

The history of the semiconductor industry is littered with attempts to couple programmable logic and application-specific blocks to create ASIC killers that can also outmatch ASSPs. Almost all of them have been blighted by Goldilocks issues that led to their eventual failure. Gavrielov is to be applauded for trying to find a route to higher growth for Xilinx. But just wishing for the market’s economics to change won’t make it so.

Slicing and dicing the European chip business
By Chris Edwards

11 April 2008

When former STMicroelectronics R&D director Jo Borel tried to convince the French government that it should encourage Europe's three largest chipmakers to merge, he almost certainly didn't have in mind what ST and NXP Semiconductors now plan to do. They are not merging the entire companies, but taking the wireless business units and glueing them together.

The argument used for the merger is not all that dissimilar to Borel's: it's all about scale. Borel wanted Infineon, NXP and ST to team up to be big enough to build and operate a leading-edge fab - it is something that is only worth doing if you are selling billions of dollars' worth of chips every year out of that facility. Not able to do that on their own, the three companies expect to buy wafers made using the latest processors from foundries such as TSMC.

The availability of foundry-made silicon is one reason why Infineon chief Wolfgang Ziebart has said that there is not all that much point in trying to be big for the sake of being able to keep building fabs. His view is that companies will specialise and do whatever they can to be in the top three of their chosen market. Infineon has been bulking up in wireless recently, thanks to its purchase of a business unit that was only briefly part of LSI when that company bought Agere Systems.

The move by NXP and ST is on a larger scale, creating an as-yet unnamed joint venture that is comfortably in the top-three wireless silicon makers and around twice as big as the next largest supplier. According to iSuppli, that will be Infineon once the deal is done. The German company is at the head of a line of $500m to $1bn suppliers. The ranking switches a little if you look at it from the perspective of baseband processors - the single most important segment in cellular wireless silicon. ST lies at number three, NXP at five.

According to Francis Sideco, senior analyst for wireless communications at iSuppli, Mediatek is currently number three behind Qualcomm and TI. ST is at four and NXP at sixth: separated by Freescale Semiconductor.

Sideco agrees with the bosses of NXP and ST that scale matters in this business. For Sideco, the turnover for a long-term survivor in the wireless space is in the $3bn to $4bn range. JV had sales of a little under $3bn in 2007, according to ST president Carlo Bozotti. TI and Qualcomm are turning over more than $5bn, according to figures from iSuppli quoted by ST in its analyst call. The merger will put some distance between JV and Infineon, according to iSuppli's number.

Just glueing business units together is not necessarily going to keep JV at number three -all too often these deals are less than the parts, let alone the sum of the parts. But the overlap between NXP and ST's wireless units is surprisingly low. NXP is good at standard products and basebands; ST has specialised in doing custom jobs for major handset makers.

Assuming the merger is successful, the deal could be the catalyst for a wave of similar deals in this business. Or it could presage a wave of price cutting as the mid-sized players try to make sure they can maintain a toehold in the market while the big fish attempt to tie up all-you-can-eat deals with the handset makers?

One thing that counts against there being a series of consolidation deals is the credit crunch. Private equity's foray into the chip business did not last as long as the private equity firms expected. And the few that are in the hands of the financiers are too heavily loaded with debt for comfort. That helps explain why the deal between ST and NXP looks a little odd.

If you just took the two wireless groups and put them together, you would expect ST to have the larger share, but not by much. Instead, ST is going to own 80 per cent of the JV and will pay $1.5bn for the privilege, all out of the company's own cash reserves. The company does not seem keen to repeat the Numonyx experience of trying to raise loans to fund the deal.

It begs the question of why NXP did not just sell the whole wireless business to ST: it would raise more cash and cut NXP's heavy debt burden. But, NXP chief Frans van Houten claims the company is in the deal for the long run, even though it has a piece of paper that explains how NXP might cash-out of the JV. NXP is presumably hoping that the value of its investment will rise as the supplier count in the wireless-silicon business falls. Margins in the wireless segment are lower than the rest of NXP's operations, so the company may feel that it can make a better return by hanging onto a small part of that business rather than selling it all now.

It may be that, without the pressure of the debt, NXP would have soldiered on alone -and tried to stay in the top five by continuing with the kinds of small deals the company did with Silicon Labs and Glonav.

Most of the other companies in the top ten are not in that position. The wildcard is Freescale, which is saddled with the same kind of debt-heavy position as NXP. But who would try to buy Freescale's handset silicon group? The company is not in the position to do the reverse. It might take the kind of chaos that has forced the memory makers to consider shotgun weddings to sort out what happens in the remainder of the top ten.

Bulk silicon is here to stay
By Chris Edwards

20 December 2007

The announcement this week by Toshiba that it will extend its research collaboration with IBM and other chipmakers in the alliance run by Big Blue underlines the key trend for the 45nm and 32nm generations of silicon. It’s still all about the cost.

Toshiba already works with IBM and Sony on silicon-on-insulator (SOI) processes. The latest deal extends that to bulk CMOS technology: the type of process used by most chipmakers in the world. It has its problems but it has one big advantage: it remains the cheapest option open to them.

Very few chipmakers outside the AMD and IBM alliance have the appetite to spend more on wafers by going to SOI. And they tend to devote SOI to high-end, high-speed processors where the margins are fat enough to absorb the cost. If the industry decided on a wholesale move over to SOI, the price of the wafers should come closer to traditional bulk-silicon substrates. But no one is keen to start that avalanche: they want to find ways to stay with cheap bulk silicon as long as possible. Even technologies being trialed on SOI today are likely to move over to bulk silicon.

Take the FinFET, for example. Practically all of the research devices shown to date have made use of SOI substrates. However, according to NXP Semiconductors, there is little to stop chipmakers from putting the thin fins onto bulk wafers. It might mean a change to the behaviour of the transistor but the main improvement the FinFET brings is the ability to wrap a gate electrode almost all the way round the transistor channel. Interactions with the substrate should still be much lower than they are today with conventional planar transistors and, by deploying them on bulk silicon wafers, the cost should be lower.

The same considerations are driving the quest to find a way of making metal gates using the same materials for the NMOS and PMOS gates as far as possible. Chipmakers can absorb the extra cost of different barrier layers here and there for the two different types of transistor. But to have to make different gate stacks pushes up the cost too high.
Because most of the market for 45nm and 32nm parts will be in low-power but cheap, high-volume applications, the chipmakers can take materials with properties that are less than ideal. But they cannot afford to put in the long-winded process steps that having two metals implies.

Faced with the cost and volume advantages that bulk silicon now seems set to enjoy for at least the next five years, the question is whether even backers of SOI will try to push as much production as they can onto the cheaper, bulk option.

Lonely at the top
By Chris Edwards

12 November 2007

Intel is spending a lot of time telling people how much of a breakthrough it has made in introducing metal gates - reversing a trend that Intel's founders began back in 1967 when they developed the first silicon-gate process. The Penryn processors are not just 45nm processors, they are “45nm hafnium-based high-k metal gate” chips.

The processors have already entered production: reverse-engineering specialist Chipworks is etching away the top of its sample right now to see what lies underneath. So the 45nm generation is underway. And, based on the combination of metal gates and 45nm, Intel looks to be way out in front. But it's not necessarily the case. Foundry TSMC has started running customer wafers through its most advanced fab. We can expect the chips from those wafers to go on sale soon. Matsushita has already got manufacturing runs of its Uniphier chip.

Where Intel does seem to be going it alone is on metal gates. So, it should come as no surprise that the company is focusing attention on that part of the process. However, metal gates are not necessary for making 45nm devices, only those that are likely to burn a lot of power, such as PC processors. For those operating in the low-power segments, the metal gate is something of a luxury. And it's a luxury that a lot chipmakers are avoiding, at least for this generation. Only IBM has said it will use them and it has yet to show off any chips made on a 45nm process.

Although the combination of metal and a dielectric with a higher dielectric constant than silicon dioxide stops current flowing out of the transistor out through the supposedly insulating gate, that form of leakage is only one small part of the overall power loss. In a high-speed chip such as a PC processor, most of the leakage is down to the transistor passing electricity even when it is supposedly turned off. These days, transistors don't so much switch between on and off; they go from on to less-on. As a result, most of the changes needed to deal with leakage have to happen during the design phase. And, those changes cut gate leakage as much as the other, more bothersome form: subthreshold leakage.

So, other chipmakers do not share Intel's enthusiasm for metal gates. They don't want the extra cost of shipping in sophisticated and slow equipment that lovingly deposits the novel high-k dielectric one atomic layer at a time. Even Intel may not choose to use metal gates on the products it plans for the consumer-electronics market.

TSMC last year said it was developing a metal-gate process for 45nm but, as the process got nearer to production, found that a more conventional gate structure was all that it needed. The foundry is looking at the metal gate option seriously for the 32nm process and will, if a customer really wants it, think about putting a metal gate on the existing 45nm process. But they are really going to have to want it.

However, Intel's move won't shove up the cost of making Penryns that much, it seems. In comparison with previous desktop processors, the die is pretty small. The penny you see sitting on a wafer full of Penryns in the publicity shots gives a very good idea of how small the chip is. That means Intel can get a lot more onto a wafer, reducing its overall production cost. This, most likely, more than offsets the increase in cost caused by the use of the more complex metal-gate process. And it gives Intel a potentially strong lead in any price war that it might conduct with AMD to push the advantage it currently holds. There are other companies with 45nm devices, but AMD is not one of them.

The fat controllers
By Chris Edwards

09 November 2007

Microchip Technology, the company that stole the lead in 8-bit microcontrollers from Freescale Semiconductor (back in the days when it was part of Motorola), is now going to have a go at the 32-bit market. The company has licensed the M4K core from MIPS Technologies, rather than give any of its own architectures another workover. Two years ago, the company turned its digital signal processors into general-purpose microcontrollers to try to break into the 16-bit market. Now it has another architecture to push.

This is something that Microchip has pulled off once before. But only once. If you rewind time by about 12 years, you can look at the circumstances that gave an aggressive relative newcomer the advantage over a seemingly unassailable market leader. The problem is that, this time around, the circumstances are different. And Microchip has yet to demonstrate that it can turn success in the 8-bit market into big sales in the 16-bit sector, let alone the even more competitive 32-bit space: a market space that a number of chipmakers have made the battleground for control of the future microcontroller business. And they all started earlier.

There is very little in the MIPS M4K that cannot be found elsewhere. However, that is true of most of the 32-bit players - their cores pretty much do the same thing.

Memory cost is likely to play a somewhat larger role in the 32-bit microcontroller market. But Microchip does not have much in the way of cards to play here. Anyone who is serious about selling 32-bit microcontrollers knows that they have to offer fairly serious quantities of flash and static memory to get anywhere. That is part of the reason why the processor core itself is not all that relevant. It comes down to how much you can use manufacturing decisions to push memory costs. The quantities of memory involved mean that these devices tend to look like memory chips with a processor core and some peripherals bolted onto the side.

This is where Microchip may end up being outpaced by companies already in the market. The Arizona-based chipmaker has chosen to go with a foundry-based 0.18um process - the company does not currently have any fabs able to deal with a sub-0.35um process - whereas some competitors have decided to push density by moving to 0.13um. Infineon Technologies in particular has a tactical advantage here as the company has had a working flash module on 0.13um for several years. The foundries are catching up but, when it comes to pushing costs on what are becoming mature processes, the integrated device manufacturers tend to have more room to manoeuvre on price. Many of Microchip's competitors are sticking with 0.18um for the moment, so Microchip is not that far behind in terms of process. But neither is it ahead.

And the company is coming into this market well behind Renesas, NEC and Freescale as well as ARM licensees such as Texas Instruments. And both microcontroller-oriented ARM cores - the ARM7 and the M3 - have picked up wide support among other chipmakers.

ARM has a distinct advantage if you take social factors into account. A lot of people who have migrated to 32-bit from 8-bit are picking ARM because it looks to be one of the dominant embedded-processor architectures of the early 21st century, similar to the way that Motorola dominated the 1980s and 1990s with its 68K family. Those social factors are hard to ignore - engineers are well aware that there are plenty of job ads for people with ARM experience. ARM's position in the mobile-phone and handheld market is paying off well here.

By opting for MIPS, Microchip does not have to compete head-to-head with so many players. But neither does it get the opportunity to pick up on customers who have embraced ARM because there is plenty of competition and plenty of engineering talent around for it.

It makes it look as though Microchip's position is more defensive than aggressive. If it cannot stop people moving to 32-bit, at least get them to migrate to a 32-bit micro supported by the same family of tools. But it's hard to be more optimistic about Microchip's chances than that, especially given the company's slow progress with non-8-bit products in the past.

History does not look as though it is going to repeat itself in the 32-bit microcontroller space.

From fab to fabless
By Chris Edwards

09 October 2007

The decision by Atmel to shutter its fab in North Tyneside should not come as that much of a surprise. Late last year when the company announced that it was pursuing a fab-light strategy, the plan then was apparently to sell the fab as a going concern.

Atmel said it planned to negotiate a supply agreement with the purchaser to continue to get products made at the plant, suggesting that a sale that would keep the plant intact was seen as feasible. But, it was clear that the emphasis was on Atmel reducing the amount it spends on capital – and that meant simply getting rid of most of its fabs.

In the end, the deal that turned up was for the equipment, not for the fab itself. TSMC, which continues to ramp up capacity both at the leading edge of production and on older processes, decided that it wanted the kit. Atmel will not separate the values of the deal for the land and buildings sold to Highbridge Business Park and for the equipment hived off to TSMC, but the bulk of the $124m is likely to be for the fab tools.

When it bought the land and buildings from Siemens in 2000, at the peak of a semiconductor boom, Atmel paid around £65,000. The company later sold 40 per cent of the land it had acquired to Highbridge, the operator of the neighbouring Cobalt Business Park, for around £8,000. It seems likely that, even allowing for rampant inflation in property prices, most of those millions will be paid by TSMC, as the fab shell itself is likely to be poorly suited to most of the tenants who have set up shop at Cobalt.

A collection of offices and call centres could be the monument to foreign inward investment in electronics manufacturing in the north-east of England, and an unfortunate one. The North Tyneside was ill-starred from the beginning, through no fault of the people who worked there. The plant was conceived amid a chip boom and then shut by Siemens when it became clear that the market for its memories had disappeared. Siemens itself got out of the memory-chip business entirely by hiving off that operation in the form of Infineon Technologies. The process has now gone one stage with the separation of Qimonda from Infineon.

Atmel bought the fab during the following, and even bigger, chip boom. Then the talk was of how the company was going to be able to buy enough production equipment to fulfil its flash-memory supply deal with Siemens, part of the package that enticed Atmel to buy into North Tyneside. Within months, there would be no problem getting hold of fab equipment – the question was how to sell the parts made using it.

However, Atmel did begin fitting out the plant for microcontrollers. But, structural changes in the industry finally put paid to North Tyneside. Chipmakers want to move production to foundries where they can and take capital depreciation off their balance sheets. Coupled with that is the tendency to cluster manufacturing – fabs of North Tyneside’s capability are springing up halfway around the world in China.

The problem is that chips have a high value and are light – the only things that stick when it comes to inward investment are products that are too heavy to ship economically around the world, such as cars. Atmel talked warmly of “sticky inward investment” when it bought into North Tyneside. People should perhaps talk more of “heavy inward investment”. In reality, the only electronics products that qualify for the “too bulky to ship easily” accolade are business computers and big-screen TVs.

Let your fingers do the talking
by Chris Edwards

10 January 2007

During his keynote at Macworld, Apple CEO Steve Jobs went out of his way to put down the stylus as a way of getting things done on a handheld machine. Fingers are all you need for the iPhone that debuts in the US in June, apparently. But that might prove to be the wrong approach for a device that will cost hundreds of dollars even with a phone contract.

The iPod user interface works with just a couple of buttons - easily replicated on a touchscreen - because you never need to put anything into it. Nothing is more than a couple of clicks and scrolls away. It would be a total pain if I actually wanted to edit a track name or a few contact details.

If I'm typing, I like to be able to see what's actually going to appear. A touch-keyboard designed for fingers is going to consume most of the iPhone's screen, even just using the multipurpose numeric keys of a standard handset. Having most of the screen taken up by a virtual keyboard works for satellite-navigation systems as you only have to enter one or two lines of an address. It isn't going to fly for anything but the shortest text message or email.

As an iPod that can make calls and surf the web a bit, the iPhone design looks good. Very good. It at least cuts the number of bits of hardware you have to tout around. Then again, the 8Gbyte Nano is barely noticeable in a pocket. For the equivalent of £300 plus a lengthy mobile-phone contract, I'd be happier with something that was going to be usable for writing and emailing on the move.

On the other hand, I would be happier if a handheld computer was based on a variant of Apple's OS X than the current offerings out there. For too long, the sales pitch has been on adding more and more features to handhelds at the expense of usability. Not only are the features hard to use, they rarely work as advertised.

I ended up ditching a Palm LifeDrive and replacing it with a Windows Mobile-based HP Ipaq because the Palm software was so, so bad. A web browser that is barely able to handle ordinary websites was just the beginning. Then there was a WiFi implementation so rubbish that the free network at MIT sent the device into a reboot panic. But the real pain was that Palm changed Graffiti, its handwriting system, to make life easier for new users. Anyone who had learned the old system found that not only were those gestures no longer worked but that the new system was slower and less reliable.

It is a testament to Microsoft's doggedness at plugging away at something until it finally begins to work that the Ipaq turns out to be a far better machine and, ironically, has a handwriting mode that works better for Graffiti users than Palm's replacement.

Even so, the Ipaq is hardly convenient to use. Just getting it to connect to a WiFi network initially seemed to involve an incredible number of settings, all in highly obscure places. Then there is the small matter of having to run the setup software under emulation on a Mac - a little reminder of how corporate vanity conspires against computer companies' desires to make money. However, the Ipaq's little foibles have nothing to do with its use of a stylus.

It is possible that the iPhone has types of gesture recognition that goes some way to avoiding the problems of using a big virtual keypad - we just haven't seen them yet - but the user interface is at odds with the positioning. In an ebook reader or an iPod that makes calls, that multi-fingered interface would be a real winner - but those would be much cheaper devices.

In short, don't expect the iPhone to be a runaway best-seller: it seems to have the wrong user interface for what it is meant to achieve and that won’t help long-term sales.

LSI Logic and Agere sew up storage as they look for growth elsewhere
by Chris Edwards

4 December 2006

The arrival of serious amounts of private-equity money in the semiconductor business has clearly focused minds. Decisions that were unthinkable before, largely for reasons of corporate hubris, have suddenly become easy to make - better to make the changes yourself as a board of directors rather than have someone come along and tell you what to do.

No-one wants to admit that private equity has convinced them to face hard truths, but with the investigation of so many companies behind the scenes by financiers, it's hard to see what other factors could have been such efficient catalysts for change. But, there is a bigger trend that lies behind the decision by companies to either strip out businesses they no longer want - as with International Rectifier's recent move to sell its power transistor operation to Vishay - or to join with other similarly sized players.

The decision of LSI Logic and Agere to combine operations in a 52/48 per cent deal is not just two former custom chipmakers huddling for protection but reflects several aspects of the current environment for chipmakers.

Some claim that private equity is interested in the semiconductor business because "the cycles are over". Only the most deluded executives can think this is the case. As Xilinx Wim Roelandts points out, the timescales in chipmaking are so different to those in the consumer markets that now dominate the spending on semiconductor components, cycles are inevitable. It's a cyclical business, always has been and seems likely to be so for the forseeable future.

No, the real reason is that companies have found that they have the choice to concentrate on only a few markets rather than try to build scale from whatever markets they found themselves to be in just so they can fill their fabs up. The fab-light strategy pursued by companies such as Agere means that they do not have to worry about whether the next $5bn will come from to finance a 300mm megafab. They can just turn up at the door of one of the foundries and have the products they really care about made there.

CEOs of several of the major chipmakers pointed out in a panel session at Electronica that they have to be at the top of the list of suppliers in each market they serve, or expect to lose the money they spend on R&D to stay in that market. The message is: if you're not number three at least, get the hell out of that business.

However, former integrated device manufacturers (IDMs) are not quite ready to specialise in just one market. For the new LSI Logic, that would be a simple choice: storage electronics. Top disk-drive maker Seagate is Agere's number-one customer, accounting for a whopping quarter of the chipmaker's sales. At LSI right now, Seagate is hovering around the 10 per cent mark. Combined, the company will have close to $1.3bn in sales of chips to the storage sector alone: more than a third of the company's total revenues.

Although the company has weaker positions in consumer electronics and Agere's other specialism, communications, the stated intent is to use the larger operation to try to build scale in just those markets, while the older broad-based custom-chip operations are left to slowly wither away. Other parts may be sold off once an integration team has worked out how the two companies will really fit together. Talwalkar has not ruled out selling chunks of the new company off, although most of it was accounted for in a presentation to analysts made shortly after the announcement of the merger went out on the wires.

The handset baseband business is one of that could ultimately face a sell-off, although it fits into the consumer category that LSI wants to push. As cellphones mutate into personal entertainment products, it's not a bad fit. And Agere president and CEO Rick Clemmer was keen to stress its position in Samsung, providing two-thirds of the Korean giant's 2.5G baseband needs. However, baseband selection is not an area where handset makers expect to spend a lot of time in the future, making it easier for companies who can sell applications processors alongside basebands. It's not clear what LSI will have to offer in terms of applications engines for phones: companies such as Texas Instruments, Qualcomm and Freescale Semiconductor are better positioned.
In contrast, the networking business of Agere is likely to stay. It forms one of the three legs of LSI's strategy for the combined operation alongside storage and consumer products. As a former arm of AT&T and Lucent, communications infrastructure runs through Agere like the writing in a stick of rock and the company continues to have significant business in this area, although it was badly hit by the recession of 2001.
As president and CEO of LSI, former Intel executive Abhi Talwalkar has demonstrated that he is prepared to make major changes to a company. He reversed the decision to spin out the storage systems business of LSI shortly after taking over. Earlier this year, he decided that it was time to sell the company's flagship chip-production plant and not just go fab-light, as Agere has done, but fabless. And he killed off the short-lived but expensive-to-develop RapidChip custom-IC project. It's easy to see Talwalkar, in charge of a larger LSI, deciding to sell the bits that no longer fit.

Mixed signals for chipmaking business
By Chris Edwards

Statistics abound in the world of chipmaking. Some of the most useful are those prepared by Semiconductor International Capacity Statistics (SICAS), a body set up by the biggest chipmakers to track how much of the world's production capacity is being used. With its copious graphs and tables, the SICAS quarterly report provides a window into the volatile world of silicon production.

The latest figures, for the first quarter of 2006, show that the current boom in the sector differs from previous upswings in some interesting ways. They tend to indicate that the market is not overheating and that the current growth spurt has some way to go. However, they also provide support for those who believe that the latest round of upgrades to the sales estimates for 2006 were misplaced.

Encouraged by strong sales in the year so far, the World Semiconductor Trade Statistics (WSTS) organisation upped its forecast for the year. Instead of sales growing 8 per cent during 2006, WSTS now thinks they will increase by just over 10 per cent. In contrast, analysts at JP Morgan thought that a slip in the April figures, largely due to slow sales of PC microprocessors, indicated trouble ahead. The firm cut its forecast for 2006 from 11 per cent to 9 per cent, believing that the market has hit its peak.

Even after the April figures were in, the Semiconductor Industry Association joined WSTS in boosting its forecast for 2006 chip sales from 7.9 per cent to 9.8 per cent. The latest SICAS figures do not show what happened in April, but they do indicate a slowdown was underway during the first quarter. They also show that the usage of fabs in the last few quarters has been much more uneven than is usually the case in a strong market.

SICAS splits its figures into groups of processes. The one that most people watch is the leading-edge bucket. For SICAS, that means anything with minimum feature sizes below 0.12µm. That means the 0.11µ, 90nm and the brand new 65nm processes. Typically, it's unusual to see these at anything less than 90 per cent of capacity. Fab owners do not like to invest lots of money in equipment to see it go idle, so they try to track the industry as much as possible.

Fab owners will normally increase capacity in response to upswings and, because of the delays in getting extra capacity online, they normally lag actual demand. At the same time, the companies selling  chips like to take advantage of the lower cost offered by doubling the density of circuitry on a piece of silicon that the move to each successive generation offers. So, they queue up to get on the new processes. The result? Capacity is pretty much sold out unless the market is in poor shape. It will be no surprise that utilisation of leading-edge processes was at its worst from the start of 2001 until late 2002.

Even in the downturns of 1996 and 1998, leading-edge utilisation only narrowly dipped below 90 per cent. In the bleak years after the collapse of the Internet bubble, utilisation sank close to 80 per cent. But that was still a lot better than what happened to older processes, all of which saw utilisation slump to less than 70 per cent and, in most cases, less than 60 per cent.

In an upswing, utilisation tends to converge in the high 90s. You can normally tell when this market is overheating when just about everything is more or less sold out. You could see that happening in 1995, 1999-2000 and, most recently, in the first half of 2004. The last cycle ended when companies noticed they had built up way too much inventory and cut back on buying chips. So utilisation of everything, bar the leading-edge stuff, quickly dropped back to the mid 80s, slowly climbing back during 2005.

Although one or two foundries have remarked on how they have sold out of capacity on some of the older processes, the industry in general has not seen the same effect. And, in the latest quarter, after utilisation nudged 90 per cent for most proceses, the number slipped back to the mid 80s again. Thanks to the lull in microprocessor demand, leading-edge utilisation also took a slight knock, dropping two percentage points.

If that situation continues it means that chipmakers will find it difficult to keep prices up, which will dampen growth. The move to 65nm may give chipmakers at the high end the opportunity to charge more per chip, which will push the average higher, but the effects of that move will not be seen for some months. And the big buyers are good at watching the markets to see who is vulnerable to a bit of price pressure. Unless demand begins to approach supply, the bears will have a better chance of being right than those who just upped their forecast for 2006.

Scott McGregor and the fabless myths

By Chris Edwards

2 June 2006

Sometime in the 1990s, the CEO of Cypress Semiconductor, TJ Rodgers, claimed: "Real men have fabs." His argument was that, although some companies had seemed to be doing OK by renting space at other people's chipmaking plants, to get real control over what you are doing you need to have a fab you can call your own.

Until recently, most companies with fabs agreed with Rodgers. They saw themselves as having the keys to unlock Moore's Law. To put yourself at arms length with the technology that drives electronics would be commercial suicide. But, in the years that followed Rodgers' claim, a growing number of companies have cheerfully bought processed wafers either from old-style integrated device manufacturers (IDMs) - companies like Cypress - or from specialist 'pure play' foundries. And they have not done too badly. This is what Scott McGregor, president and CEO of fabless chipmaker Broadcom sought to tell executives from other companies in the industry at the recent International Semiconductor Executive Forum in Munich, organised by the Institution of Engineering and Technology (IET) and the Fabless Semiconductor Association (FSA).
McGregor used to run Philips Semiconductors, one of the biggest IDMs in the world. But Broadcom has convinced him that everyone should go fabless. Maybe not tomorrow, but sometime. He decided to engage in a bit of mythbusting, listing ten statements that people have used, apparently, to argue against the fabless chipmaking model. Some of them are familiar; some seemed to be strawmen designed to be dismissed easily and quickly.

McGregor's first myth was: "fabless companies can't achieve significant revenue growth". He plotted the sales figures of fabless vendors and IDMs since 1995 and showed that, on average, the fabless companies grew more. However, as fabless companies started off small at the start of that period, anything less than stellar growth would be a victory for the IDM model.

The second myth declared: "fabless companies can't achieve acceptable profit and gross-margin levels". This one is less of a strawman argument. The argument on the side of the IDM is that you get to keep the money that fabless companies lose in the profit margin they have to pay to their foundry suppliers. McGregor put up figures claiming that gross margins on a basket of IDMs and fabless suppliers he selected were 8 per cent better on the fabless side: 50 per cent for those without their own plants; 44 per cent for those with. McGregor did not go into detail on it at this point, but a lot of this has to do with fab utilisation - a theme I will return to often. IDMs have to pay for their fabs whether they are busy or empty. When fabs are empty, they cost a lot of money. That can put a serious dent in your gross margins.

The third myth was "fabless companies can't multisource". It seemed a strange myth to me. Was there something in contracts with foundries that said fabless companies can't go to the competition? Xilinx has been using at least two foundries for many years. Similarly, McGregor said Broadcom has made ample use of different foundries for the same products to help guarantee supplies. "We can tape out the same product to any one of these guys," he said, flashing up a list. "If this foundry is full, we can shift the load. If there is an earthquake at one, we can tell customers what we can do to restore supplies".

Fourth on McGregor's list was: "fabless companies can't develop integrated mixed-signal circuits". Again, a rather easy one to knock down for anyone. Plenty of companies selling products with large proportions of analogue circuitry on them do not have fabs. McGregor said this myth came about because the IDMs have specialised, exotic process such as BiCMOS and silicon germanium bipolar that they don't offer to other people. Curiously, the same day, Infineon Technologies said it would offer access to its more exotic processes.

"They say they can do really cool stuff, unlike the fabless guys. People say you need those fancy processes or you can't do high-speed chips," said McGregor. He countered by claiming that Broadcom had done advanced mixed-signal chips in ordinary CMOS processes available at any of the top foundries. Examples were chips that could support 10Mbit/s interfaces that were implemented using 130nm CMOS from foundries. "In working with the foundries, if you have a good engineering team, there is no limit on what you can do," he claimed.

The fifth myth cited by McGregor was: "fabless companies can't develop innovative products on time". The argument from the IDMs here was that owning a fab got you engineering samples faster from new processes. Again, the Broadcom boss declared reality is different - the foundries are, in some cases, providing access to advanced processes before many of the IDMs. In reality, you have to have good connections with the fab in either case to get samples through quickly using what the industry calls 'super hot lots'.

That fabless suppliers can't get traction from investors was McGregor's sixth myth. "They can deal with VCs but when it comes to Wall Street it's different," he said. This was relatively easy to deal with. Even I was thinking of the way that shareholders tend to dislike capital-intensive businesses with highly cyclical returns. "The IDMs suffer on Wall Street," said McGregor. However, some of the best performers on the graph that he showed did own fabs. Above a line drawn as the sector average were companies such as Maxim Integrated and Microchip Technologies. The link between those companies lay in small, specialist analogue or mixed-signal products. And very, very old fabs. That is something that has not escaped the other IDMs.

Myth number seven was similar: "fabless companies can't gain the trust of the major OEMs". There was a hint of straw about this myth too. Major OEMs dislike changing to companies with little or no track record, unless they get a good deal that makes sense for them. After a while, they get used to dealing with new kids and start treating them as long-term suppliers. This myth was quickly pushed aside. McGregor said this is no more difficult for a fabless company than for an IDM.

The eighth myth was more interesting, that "fabless suppliers can't get enough product during allocation". This myth had a particular resonance at the conference. Supplies were getting scarce and the managers there were thinking of previous booms, such as in 1995 and 2000, when customers had to be told that there were no more chips to be bought. In 1995, the allocation problem was particularly severe for fabless companies. There were stories of fab owners selling access to wafers twice over. Companies that thought they had guaranteed access to production found their wafers had been sold to some other company. It was at that time that the first batch of big fabless companies found themselves getting tied into deals that netted them 10 or 20 per cent of a fab in new consortia.

McGregor himself did not allude to 1995 but to the more recent bubble of 2000, when he was at Philips. "I don't have a detailed statistic on this but in 1999 and 2000, the toughest years of the last capacity crunch, we had no allocation to some customers during 2000. But the fabless companies did get their wafers," he said.

Myth number nine stated: "fabless companies will outgrow their foundries". The idea behind this one is that some fabless companies will get so big, they might as well build a fab because they will be able to fill it just with their own products. "I often get asked that by investors," claimed McGregor. He cited an 11 per cent annual growth rate in foundry capacity as one reason why there would always be enough rented fab space to go round. And declared that Broadcom would not build a fab even with sales of more than $10bn - a level more than healthy enough to justify sole ownership of an advanced fab by an IDM. "We can remain healthy and growing as long as the foundry industry is healthy and growing," he said.

And the tenth myth claimed: "fabless suppliers can't gain access to the latest and greatest technology". This was more or less like the fifth myth. However, McGregor cited Taiwanese foundry TSMC as the prime example of why IDMs did not have all their own way. "TSMC leads the ITRS Roadmap," he said, alluding to the document prepared by research consortium Sematech that is regularly updated to show when new processes are expected to come through. If you look at the schedules from the major fab owners, TSMC is up there with the leaders, but is behind some and ahead of others.
Having dealt with the myths, McGregor looked at the core problem for IDMs: "If you have a fab and it's only 80 per cent loaded, you think a lot about how you are going to fill it." That means selling chips for less than they would like, simply to get the utilisation levels in the fab up. Eighty per cent is not a good number - it is the kind utilisation level people in the industry associate with downturns. "The IDMs imagine they have a margin contribution but they don't. They should be realistic about that," concluded McGregor.

The widening foundry gap

by Chris Edwards

18 May 2006

There is a lot of talk at the moment about the haves and have-nots of the foundry world. People worry that the market is dividing into a top-tier oligopoly that controls all the leading-edge production and a goodly chunk of the rest. And then you have ‘boutique’ fabs and a crop of integrated device manufacturers (IDMs) who have turned over some of their capacity to foundry production. That separation was inevitable and the gap is going to get even wider. It is the inevitable result of chipmaking economics.

I think there is a good chance that four is too high a number for the top tier. If you take into consideration the fact that a number of the top-ten IDMs have gone ‘fab lite’ in recent years in order to take advantage of the production efficiencies that large foundries enjoy, four is quite a high number for top-tier pure-play foundries. If scale is necessary for IDMs to invest in fabs, then surely scale is the dominant criterion for foundries. Network effects also play their part: because more intellectual property suppliers will target the biggest foundries, those foundries will get more of the designs. Any foundry that falls behind on investment or technology will start to lose that support.

We have seen in the latest boomlet that the benefits accrue to the largest player in the market. And that will help TSMC build out the capacity for the next process generations in a more aggressive fashion than its smaller competitors. Unless TSMC makes a serious mistake, it is likely to grow faster than the overall foundry market – although that latter number may be distorted by IDMs coming into the business with their older fabs.
What happens to the boutique foundries is going to be harder to predict. At the recent IET/FSA Fabless Forum, many fabless companies explained how they liked to design for vanilla CMOS because it is cheap. The boutique foundries have specialist, often exotic processes that support a narrow range of applications. If everyone goes for CMOS, what do they do? CMOS cannot do everything. And trying to stretch it too far will result in something just as expensive as the specialty process. The key for the boutique foundries is to make their offerings more attractive and easier to use for people who are teetering between CMOS and something more exotic. So much money goes into CMOS circuit design, that the techniques in other processes have fallen behind. If foundries can collaborate with specialist designers to make the most of the processes, there are places where they can overtake vanilla CMOS in terms of cost and performance.

The final ingredient is greater co-operation between boutique foundries. Right now, there are too many processes with “that little bit extra” that looks like a lock-in to any customer. Many fabless companies use vanilla CMOS because it is easier to port between foundries – it reduces risk. Greater emphasis on reducing risk in the boutique world this way may help a lot.

Keeping consumers’ music safe

By Chris Edwards

27 March 2006

Apple quickly rounded on the French parliament for daring to pass a law that demands interoperability between digital rights management (DRM) systems and media players. In trying to enact a European directive calling for the protection of digital content as agreed by the EU with the World Intellectual Property Organisation, and prevent reverse engineering of DRM and similar systems, members of the parliament tried to level the playing field a little. They passed a law that demands that digital content - of any type - bought online should be playable on any type of music player.

The decision seems to affect Apple the most because the company has refused to license its FairPlay system to any other manufacturer. Microsoft will also be affected but, because it licenses its system to hardware makers, the problem looks less serious.
Apple has tried to paint the French move as tantamount to legalised piracy. That is stretching the truth to breaking point. The position taken by the French parliament seems entirely reasonable. The consumer has bought the music; the consumer should not be restricted in what computer or portable device that music is played back on. Nobody wants to be in the situation ten years from now where they find they cannot listen to a song just because they don't want to pony up for a new iPod - maybe Creative finally got around to making a player that looks good and works better. The French government is not asking for DRM to be removed; simply for decoding software that understands the format to be able to run on any computer. Apple has no real justification for maintaining a lock over iTunes songs. It is not as if cross-subsidies are involved, although the company might regret having cut a deal so friendly to record companies in the hope of driving iPod sales if FairPlay does have to be licensed out.

To be fair to the company, its copy-prevention system is more flexible than most of the alternatives. Backing up songs from iTunes actually involves turning them into CD tracks that you can then rip to move the songs to just about anything you like. Should France target Apple, this might prove to be the company's defence for keeping FairPlay to itself: everything needed for interoperability is already in place. It's not seamless. But it would allow you to play the music on anything you like, as long as you have a copy of iTunes handy to do the job. It is easier than trying to do the same with a Windows Media file that has been encoded using the Microsoft PlaysForSure system.

The wrong structure

by Chris Edwards

8 March 2006

LSI Logic’s timing for its exit from the structured ASIC business was impeccable: the very first day of the EDA-oriented Design Automation and Test (DATE) conference. It is a conference where – as with its US analogue, the Design Automation Conference – the background concern is one of the number of chip-design starts and, therefore, the number of customers.

Over the past few years, the attention of DATE has shifted away from SoC and ASIC designs to the much larger market of field-programmable gate arrays (FPGAs). Unfortunately for the design-tool companies, it is less lucrative. The structured ASIC – effectively a return to the old gate-array business model – was designed to stop the haemorrhaging of design starts as many people found the costs of designing for deep submicron processes way too expensive to contemplate. By trading area for ease of design and lower mask costs, LSI Logic hoped that its Rapidchip products would shore up what looked to be a fading ASIC business.

The problem that faced LSI, along with its former competitors in this space, was that it would take a lot of designs to see a return on its investment. The best ASIC customer is one that needs a lot of chips: millions of them. Many projects are smaller, but there is no ceiling on what customers might order. Because of the area penalty, the structured ASIC is different. Beyond a certain volume, it does not make sense commercially. You might as well bite the bullet and go with the ASIC. That puts a ceiling on the returns any company can see from a structured ASIC project – unless a customer has the rare surprise of a runaway bestseller that it does not have the time to redesign. And there is a floor below which the structured ASIC cannot compete: this is the world of the FPGA.
Structured ASIC offers the promise of lower non-recurrent engineering (NRE) costs than a full ASIC. But it is still a lot more expensive than the minimal NRE associated with an FPGA. The difference can account for thousands or even tens of thousands of parts. The problem for the structured ASIC market is that it is piggy in the middle between lucrative volume and suppliers who have businesses that are about shipping small numbers to many customers cost-effectively. No matter how effective a company like LSI could be in implementing structured ASICs, its market would always be limited.
There will be a place for structured ASICs, but it will continue to be a specialty market for those who cannot go with the options of standard-cell ASICs or FPGA. Maybe they have low volume but care about reverse engineering; or they need lower power and cannot afford an ASIC. But, again, there are only so many of those customers around. So, don’t expect the big names to hang around in structured ASIC – it is a business for the niche supplier.

by Chris Edwards

11 January 2006

At the Consumer Electronics Show in Las Vegas last week, Sony unveiled its latest attempt at making an ebook reader, a product that a number of vendors have had a go at over the last 15 years, without much success. This particular iteration in the development of the ebook reader – Sony is not the only vendor to have a design coming through – may actually see this category begin to ship decent volumes.

The story behind the latest crop of ebooks is all in the display and what that means for what could be one of the highest volume niches in portable computing. Companies have been striving to find a killer appplication for handheld computers and keep coming up short. It’s not just because a lot of the software sucks. They have lacked the two major requirements in any device that seeks to replace paper: the ability to run off a couple of AAA cells or maybe even coin cells not just for hours but for weeks; and a display that does not make your eyes water after a couple of hours.

The story is really about electronic paper: a display that keeps everything visible even when the power has been turned off. That makes for dramatic improvements in battery life for the computer behind the display. It only has to wake up to do the equivalent of a page turn. It can then have a good long snooze, with the merest trickle of battery keeping a clock going or to sniff the airwaves to look for any updates for content you have downloaded. The designs of the 1990s only had liquid crystal displays (LCDs) to work with. They would get no battery life advantage over other handhelds and the contrast ratio was nothing like paper: a non-starter in an environment where displays are still too tiring to read for any length of time.

In the latest crop of products, electronic paper looks as though it might finally be viable in mass-production devices, at least from a reliability and readability standpoint. There is still some way to go before the displays genuinely rival paper, but it is now possible to see a path from where ebook readers go from being curiosities to the one device everybody uses.

At several hundred dollars, Sony’s reader is way too expensive for the market the company expects it to serve. But, this is the wild and wacky world of electronic gadgets. This is largely a concept design that is only just about ready to make it as a niche product that a few will fork over several hundred dollars, or rather, 40 000 yen to get hold of. In the Netherlands, a Philips spinout, Irex Technologies, has a more or less equivalent design that the company has decided to aim at business users - namely subscribers to expensive information services. Those early adopters might be enough to give display makers such as E Ink the necessary experience with mass production to start bringing down costs to where people will think, “what the hell, I’ll get one”.

I don’t see a price north of $50 being viable for an ebook reader if these things are to ship millions a year. This is not an all-singing, all-dancing PDA-phone that can command anywhere near a higher cost. But the nature of the electronics inside one of these things should make a sub-$50 price target, even sub-$20, possible quite quickly as long as there are no gotchas in the large-scale production of the display technology. The device could well piggyback off a phone for Internet access and other more complex functions. The ebook reader does not need a lot inside it. But the display has got to look right, it has got to be light and it has got to be cheap.

In search of scale
by Chris Edwards

20 December 2005

Scale, or the lack of it, in Philips Semiconductors’ operations was apparently a key consideration in the decision by its parent company to separate it out. With a new legal structure, the semiconductor business will be freer to pursue a merger with another large chipmaker, potentially reducing Philips' stake to a minority interest.

According to chief executive Gerard Kleisterlee, Royal Dutch Philips wants to reduce the volatility of its businesses as it turns itself into a lifestyle and healthcare company, and chipmaking is way too volatile for comfort.  In other technology businesses, such as flat-panel displays, the company is progressively selling off its shares. Once Philips Semiconductors has merged or floated it seems likely that Philips will gradually relinquish its ownership.

The management of Philips, both parent and chipmaker, echoed the statements that have been made by honorary chairman of STMicroelectronics Pasquale Pistorio over recent years that the industry was set for massive consolidation. He wanted ST to be among the consolidators, arguing that scale was essential to compete in an industry where margins would be under continuing threat.

Although Pistorio said he wanted ST to be one of the first movers in any consolidation process, the process has taken longer than expected. No deal has emerged that would allow ST to pursue a large merger. Until maybe next year, once the Philips paperwork has been sorted. Even then, the Philips move may not open the floodgates in M&A among chipmakers. ST has to be top of the list of potential suitors yet another spinout, that of Freescale Semiconductor, in which ST was thought to be interested, remains independent.

That Infineon Technologies seeks a flotation for its memory operation may indicate that there is no sufficiently large player with an interest to pursue a merger with the soon to be independent memory maker. That is in another business that seems even more prone to massive consolidation, as memories are more like interchangeable commodities than many of the devices made by Freescale or Philips.

You have to go back and look at whether consolidation through mergers and acquisitions will actually work for large chipmakers. The portfolios of the large players overlap significantly. Any merger would result in tens or hundreds of hard choices over which products would get investment and which would be left to wither. Then look at the competitive scenario. Scale is important to IDMs as they have their own fabs to fill. But many chipmakers are pursuing foundry-intensive, asset-light strategies. That relieves them of the need to have business just to fill fabs. The only bit of Philips that depends on inhouse fabs is the multi-market operation – which should be quite profitable as it relies on old, fully depreciated fabs, not new and expensive gear.

The most intense competition in the target markets for companies like Freescale, Philips and ST comes from aggressive fabless players that concentrate on individual markets. These companies have shown they are more nimble and more profitable when they get the design-ins. That can keep the big players out of some markets altogether. The incumbents have some advantages when multiple vertical-market chips have to be combined into one for cost. But it does not take big-ticket mergers to get the necessary designs together. Acquiring smaller players may make more sense.

So, although a big merger would make it easier for Philips to divest itself of a volatile business, that might not be the best answer for Philips Semiconductors. Although not a solid top-ten player, a combination of focus and small-scale acquisitions would probably make for a long-term stronger position in Europe. A mega-merger with a company such as ST or the other part of Infineon might look powerful on paper, but could severely dent the competitiveness of the combined operation.

Who needs a masterplan?
by Chris Edwards

24 November 2005

In its report on the state of the semiconductor industry in Europe, the European Semiconductor Industry Association (ESIA) sees attracting chip production to Europe as vital. But who would be helped? The large local players do not seem to be in a hurry to build the megafabs that would place Europe back among the worldwide leading countries anywhere.

Infineon Technologies has committed itself to an asset-light model where much of the manufacturing is handled by the Singaporean foundry Chartered Semiconductor. Philips Semiconductor and STMicroelectronics have shares in the Crolles2 fab but are not in a hurry to build a lot more for the most advanced process nodes. Both are users of foundry manufacturing, mostly in Taiwan.

ST is fitting out a 300mm fab in Catania, Italy which will expand its capacity for advanced logic devices. But close to half of ST’s advanced chips are already made by foundries. The capacity expansion made possible by the M6 fab in Catania is to deal with growth, not the share given over to foundries. ST has hinted its next fab might be sited in China, but the reasons given were that it would make business easier in the country if it had more local production, not simply tax breaks.

Infineon’s memory group, which is soon to be sold off, might benefit from the policy. Like AMD, the company has benefited from strong incentives to have memories made in Europe, in Dresden. There, significant tax breaks were granted to attract fabs to the city. The local government officials reckon that they got value for money: the personal taxes paid by the workers at the plants more than made up for the loss of corporate taxes.

However, without a strong push by local companies to build many more fabs anywhere, the overall increases in jobs would not be significant. A policy to attract megafabs from elsewhere would now have to go far beyond what China is prepared to offer to make Europe attractive. Even then, manufacturing is consolidating. So, it is a buyer’s market for tax breaks and incentives, which makes a race to the bottom quite dangerous.

The approach taken by the ESIA overall is problematic because its solutions are oriented to the biggest companies. The push for tax credits is fine and for greater centrally organised European R&D also seems worthwhile on the fact of it. But these are policies that favour large corporations with staff dedicated to form-filling. If Europe is to stay close to the leaders in technology, it needs to find ways to stop smaller companies feeling the need to move out, not the big ones. Large corporations like to perform what the economists call ‘rent seeking’. “Give us benefits or we move out,” is the unstated aim of this lobbying. The small entrepreneurs and businesses do not bother with this. They just move if they can or get on and grumble about the amount of work needed to keep up with paperwork.

Although policies to establish big research projects and big fabs look good on paper, they will do little to stimulate growth as the larger companies will simply use that money to have their own research and development funded by the taxpayer. The taxpayers include the smaller companies who actually could do with the cash they effectively stump up to help their larger competitors. The policy needs a lot less top-down intervention. The emphasis needs to be on finding out what the fast-growth fabless companies actually need as that is what will drive a broader based European semiconductor business. The IDMs are big enough and tough enough to do their own lobbying on their own.

Waiting for some service
By Chris Edwards

16 November 2005

Is the electronic gadget industry on the verge of having its current bubble popped? It is certainly showing the signs. Even websites that make a living from promoting the shiniest gizmos are displaying a certain amount of ennui at the endless stream of ill-considered me-too gadgets that could have done with a bit longer in the reliability lab.

Many of the changes in their design come down to different mixtures of function as phone makers try to work out whether they should be in the MP3 player business, or MP3 player makers wonder whether they should add more games. Aside from the race to build the smallest or the shiniest, the feature sets seem pretty well locked down.

This is the problem for a market that increasingly depends on external factors to make things happen. Digital audio players, for example, are almost unusable without a personal computer. And this is where the second of the problems comes in. The media and services suppliers are convinced that the consumer will do everything they tell them and pay through the nose for the privilege. Sony’s mistaken belief that no-one would care that a CD planted a rootkit, which could act as a shield for hackers, on their PC only serves as an example of how far vendors will go if they think they can get away with it. It was only howls of outrage that begun when it first became clear what Sony had done that convinced the company to stop. Sony BMG first issued a patch to stop the offending code from hiding as a first reaction and then, finally, said it would suspend production of discs carrying the protection code.

Restrictions that force users to only send messages to each via certain services when they use a particular mobile phone and provider does not encourage people to make more use of electronics. Although there is a constant stream of new users in the youth market, a growing number of people are getting fed up with the artificial restrictions placed on the gadgets they already have. And fed-up consumers stop buying. The smarter consumer electronics companies need to spell out to their new-found “partners” that the consumer cannot be regarded as a willing cash cow. Because, when the milk dries up, everyone will be struggling.

Eyes on the global economy
by Chris Edwards

3 November 2005

If the chip market takes a tumble in 2006, for once it will not be through self-inflicted damage analyst house Future Horizons has claimed in a recent report. The prospects for chipmaking depend, for the next year at least, purely on the economy. The trouble is, there is a lot of uncertainty over whether the global economy will continue to grow or will run headlong into recession as various problems pile up.

Future Horizons president Malcolm Penn said the economists at the International Monetary Fund (IMF) have predicted robust growth for 2006: 4.3 per cent. But, Penn has pointed out on previous occasions, economists rarely predict recessions. They just seem to happen. That is a little worrying for chipmakers who have endured another nervous year where the trends all seem to be in the right direction – inventories down, volumes up – except for the one that will make them feel much better: rises in the average selling price of chips.

On the face of it, the IMF’s projection is good news. “A 4.3 per cent world GDP growth is a very big if, given the inevitable slow down in consumer spending, but then it is the consumer that has been driving growth since the 2001 crash and we are thus not mentally able to believe business will finally pick up spending again,” said Penn.

“The IMF clearly believes that believes that business has the capacity to do this, and that the transition will be smooth. If this is the case, the chip market will boom. We have seen a semiconductor market correction in 2005 from which the industry can bounce back fast if the economy remains strong,” he added.

The semiconductor industry has got its house keeping in order during 2005, Penn claimed, such that growth in the next 12 months will be tied to demand and thus the strength of the world economy. Future Horizons has reported that three key variables for the worldwide semiconductor industry are now unusually aligned: excess inventory levels have been purged, capacity utilisation is starting to rise, and new capacity investment is moderate.

However, if the global economy slows though, the analyst says that it will bring the semiconductor market down. A global recession in 2006 would tip the electronics market strongly negative; a slowdown to between 3.0 to 4.0 per cent GDP would reduce the growth to single digits or even shrinkage if the collapse is bad.

High oil prices and consumers coming to realise that they have overstretched in the credit department will not make chipmakers too confident in the IMF’s predictions. However, the structural shift eastwards may mean that even slowdowns in the western economies will be outweighed by high growth in China and elsewhere. 2006 seems set to be another rollercoaster year.

Companies in the middle improve R&D spending picture
by Chris Edwards

24 October 2005

The R&D Scoreboard prepared by the UK Department of Trade and Industry (DTI) used to be pretty depressing reading if you were looking at it from an electronics perspective. For years, R&D spending as a proportion of revenue saw electronics hardware and software, and the related sectors, trounced by biotechnology and pharmaceuticals – industries where the UK has built up an enviable position.

Now, according to the rankings of UK and global companies prepared by the DTI, electronics R&D in the UK has not surpassed the global average in terms of spending. But it does not look half as bad as it did just five years where the trends seemed solidly down. Part of the change in perspective may be down to the growth of the report itself. The DTI decided to increase the number of companies surveyed from 750 to 1000. This meant that close to 240 medium-sized companies – those with sales up to £500m – were added to the analysis. This move has undoubtedly helped to improve the figures for electronics and IT hardware as many of the companies with the most intensive R&D spending programmes today are in those groupings.

There is a problem with including more medium-sized companies in the analysis. Their spending patterns tend to be more volatile as a number of companies in the electronics and IT hardware sector tend to move from periods of intensive R&D to lower levels of investment as they roll their technologies out, perhaps providing boosts as they grow or acquire new, less well-developed subsidiaries.

The UK figures show that, for electronics at least, a lot of the R&D spending by value rather than proportion of sales has come from inward investment recently. This demonstrates that the UK still has a rich talent base from which to draw although this investment may be more prone to a shift eastwards as more Asian engineers arrive. However, the medium-sized companies have shown they are equally willing to use offshore teams for certain projects, so the distinction between the location of the parent company is perhaps becoming less relevant.

The key things that the UK needs are a continuing flow of engineers, which is becoming more problematic as university applications for relevant courses fall, and an economic environment that promotes efficient R&D. The DTI and HM Revenue and Customs have promised to simplify the tax credits scheme and make it clearer. That should make it easier for smaller companies to boost their R&D investments.

Less is more in the venture business
by Chris Edwards

14 October 2005

A report  prepared jointly by the European Commission and the US Department of Commerce has concluded that venture capitalists (VCs) are going about it the wrong way when it comes to investing in startups and has some advice on what government could do to improve the situation. One of those things involves giving VCs more money, this time from the public purse.

If there is one thing that big VC firms are not short of, it's money. They are awash with the stuff. They have more money to hand than they can find companies to invest in. The last thing they need is an injection of cash from public sources. The idea of using taxpayers' money to subsidise big VCs is nothing short of bizarre.

The problem stated by the report is that the VCs are failing to invest in enough early-stage companies. The report does not provide direct evidence of that, but asserts that VCs have become more risk-averse. After the crash of 2001, what did they expect? A lot of VCs got burnt and a good many are still getting burned by poor investments almost forced on them by an overflow of money. The investment pile-up that surrounded WiFi and is now happening to Web 2.0 companies and Wimax specialists will lead to many more casualties. Companies that should not have been funded are coming to market with products they cannot afford to sell. Money is not the problem; it's discrimination.

It is easy for me to say discrimination is the problem. I'm not faced with the task of auditioning enthusiastic founders day in, day out when you know you have to do something with the money sitting in a bank account. A number of VCs have said some benefit might come from having fewer operating in the technology sector: it would reduce the number of companies trying to get into each niche. But there is not much that can be done about discrimination problem aside from the steady improvements to analysis processes that the VCs are making themselves. The report's recommendation that government gets involved in training better fund managers also seems a little misplaced. VCs don't like losing money anymore than anyone else, so anything that works is going to catch on fast. This is a business after all.

The report rightly says the tax and legislative environment in some countries can penalise VC investments, but the proposed changes do not readily address the early-stage problem. But there may be things that government can do to improve the level of investment in early-stage companies without directly spending tax money: provide incentives skewed to those investments. That might seem to lower the tax take later on, especially as many early-stage companies will flame out. But, if the report authors are right in thinking that VCs are missing opportunities in their current practice, there will be more successful companies as the result of the altered policy, and the future tax take goes up. The only problem is the difficult-to-test assertion that not enough early-stage startups are getting VC funding. What if the ones not getting the money today were going to flame out anyway?

Less power please, we have travelling to do
By Chris Edwards

26 September 2005

I had something of a dilemma when last packing for a trip: how much stuff did I need to take with me? The phone went but the PDA stayed at home because for this trip, it did not make sense to take it. The problem was of bulk: the PDA was small and light enough. It was the rest of the paraphernalia that went with it that caused the problem.
Taking the PDA meant taking yet another battery charger on top of the one for the phone, on top of the other bits and pieces I wanted to take with me. It sounds like I might be a candidate for a so-called converged device that does it all.

There is a compelling argument for having a so-called converged device. It will at least keep down the number of power adapters and chargers that I have to pack when going on a business trip. But there are issues with this kind of device: it is all your eggs in one portable basket.

In a recent survey of users carried out by TNS, the top requirement was not some whizzy new feature such as TV on a phone or instant messaging: it was simply for the standard facilities to last a bit longer. Two-thirds of mobile phone and personal digital assistant (PDA) users said they wanted two days' active use worth of battery life. Insufficient battery life is still a major drawback of many devices even now and many said they would not listen to music, play games or watch TV on their mobiles too often for fear of draining the battery. I can tolerate the iPod running out of juice, but what happens if my superphone leaves me unable to make a call because I did not realise how much MP3 time I had left?

In the short to medium term, battery life is not going to get better quickly. The incremental improvements in energy density tend to get chewed up by faster processors. There are a number of advances being made in power saving in processors, but all that does is focus attention on how much power audio outputs, wireless interfaces and displays use. There are limits on how low the first two can go. We can expect a step change improvement in power efficiency for displays as organic technologies take hold. The problem is that, by the time they are ready we will probably have longer-lasting fuel cells to play with as well as batteries.

Rather than a super-efficient superphone I have one request: get the power inputs standardised. Is it too much to ask for one battery charger and let that and each device sort out what voltage and amperage each needs and can sustain? Are electronics manufacturers completely unable to talk about standardising a low-speed SPI-type interface to self-adjusting power adapters? I have drawers full of wall warts that differ by 0.5V or simply by the pins they use – thank you Sony. I don’t need to have everything to converge in one device, but I would seriously consider a more expensive, one size-fits-all power adapter.

The downward spiral
By Chris Edwards

5 September 2005

The latest chip sales figures from the Semiconductor Industry Association (SIA) highlight a continuing problem that the chipmaking sector has: it has lost control over pricing. Some sectors seem caught in a deflationary spiral that even rapid advancement to smaller geometries will not help.

According to several analysts, component prices in the PC sector are dropping fast and it is a pretty safe bet that those falls will have an unequal impact across the chip supplier base.

Cellphones, where manufacturing has shifted quickly to the Far East, are also seeing rapid price falls. Some of this may be attributed to the two-speed market for phones. The biggest price pressure generally is on the cheap, commodity phones rather than the high-end phones. However, as operators migrate to 3G, they are insisting that the amount of subsidy they give to hardware manufacturers cannot go back to the levels they had when analogue switched over to GSM.

It used to be the case that, during upturns, the chipmakers could control pricing much more effectively. Prices would rise during the good times and fall back when the market collapsed. Overall, the companies would get a reasonably stable pricing environment. That picture collapsed during the late 1990s when, even during the Internet bubble, the industry suffered sudden downturns and pricing falls. After the bubble burst, prices went into free fall, recovering some ground only in the last couple of years.

The problem that the chipmakers now have is that in the good times, prices can fall. And in the bad times, when oil prices feed through to dampened consumer demand worldwide, chip prices will still fall.

The question is how much more downward pressure on pricing the semiconductor companies can take before things start to unravel. Process scaling is not what it used to be. The dramatic price reductions that suppliers can offer are based on their ability to squeeze twice as many transistors onto successive generations of chips at roughly the same cost. Various technological issues are making that simple scaling harder to achieve, which removes their traditional way of dealing with external price pressure.
We have already started to see some of the bigger manufacturers pull back from markets where margins are plummeting and concentrate on those areas where they can turn a higher profit through being the market leaders. As a result, the intensive competition between suppliers will focus on small groups of the largest players who should be able to hold price levels up, at the cost of the equipment builders and the consumers.

The wildcard remains the position of the venture-funded startup. Entry to the largest markets is becoming increasingly expensive. However, there is still more money chasing investments than there are realistic investment opportunities, especially in emerging markets such as China. The unfortunate consequence of the pricing spiral is that investors will pile into startups with big plans but flawed models. Those investors will once again see chipmaking as a low-margin business and pull out more cash than they should and starve the players with more realistic business models. In the meantime, pricing will be more bad than good for chipmakers.

A touching faith in machines
by Chris Edwards

26 August 2005

Perhaps because we know that people can be untrustworthy but machines are generally consistent in their ability to handle a situation, or fail trying, we seem to put more faith in the machines, even when the machines are clearly wrong.

In the bank, when “the computer says no” because it has the wrong data or the program is just badly thrown together, the consequences are annoying or just funny. In a car, the problems can be a lot more troublesome, if not catastrophic.

A colleague and I were being driven to an event in a car the driver had picked up just that day. Mounted on the dashboard was a state-of-the-art satellite navigation system. The driver had had long experience of dealing with computers telling him which way to go. As a former RAF pilot, he had plenty of time with ‘Bitching Betty’, the guidance computer with a synthesised voice that tells pilots “pull up, pull up” among other things as they try to complete tricky turns.

For the duration of the journey, the satellite-navigation unit would be a new form of Bitching Betty. We only hoped that the computer used by the RAF has a better sense of direction and distance.

The problem with the system was that it seemed to be programmed with a limited vocabulary when it came to describing how the next junction looked, and how far away it was. There were frequent exhortations to turn left or right in 100m. The only problem was that the machine’s interpretation of 100m was somewhat different to reality. After a while we started taking bets on how far off 100m Bitching Betty would be next time.
Before we got to that point, there was a period during which we were almost going round in circles. One key junction was no more than 20m away when the order came in to turn left in, you guessed it, 100m. So, pretty quickly, the computer was recomputing the course and we had to go around for another go. The problem was this kept on happening until the point came when we all decided that it was time to ignore the commands and just look at the map. Had we used the map in the first place, everything would have been fine.

However, before that point, there was a period of uncertainty: do we trust the computer? Do we ignore it? In this case, there were no accidents but it is easy to see situations where people suddenly turn in the road because they suddenly realise their supposedly infallible navigation systems are not up to the job.

We really need to get used to treating machines as being wrong until they can demonstrate they are right most of the time. That might also ensure that manufacturers do a lot more testing of the software to check that their systems do not simply provide misleading information.

Plugging The Gap In Engineering
by Chris Edwards

Late last week, the Confederation of British Industry (CBI) warned that not enough students were choosing foreign language or science subjects at school. According to the group’s figures, the number of 16 to 18-year-olds taking A-level physics more than halved between 1984 and 2004. Just one in twenty five students study a modern language at A level, with hardly any choosing a language that should prove useful in the future. The CBI picked out Mandarin, Russian and Spanish as important future languages.

There is a danger in forcing education to fit apparent business needs: business cycles have much faster cycles than those of education. Things that look important today often turn out to have an insignificant impact on the future. Right now, Mandarin looks a safe bet as a language for the future. But the wheels could come off the Chinese economy in the next five years. It is one of the processes of having a developing economy. Ten years ago, one might expect business leaders to pick out Korean. Twenty years ago, it would have been Japanese. Ten years from now, who knows? It could be Chinese still. Or maybe Arabic. And, although Digby Jones said on Radio 4’s Today that French has lost its importance, it is important to remember that French is not tied to Europe. Twenty per cent of the African population is Francophone. Few of the West African states will have a technologically active economy in the next ten years but 20 or 30 years out, much can change.

However, that is not to say the CBI should not be encouraging the study of languages other than traditional French and German. Aside from anything else, the study of a non Indo-European language makes people think about the structure of language itself. But language study is something that can be done, and really should happen beyond secondary education. Organisations such as the CBI do not do enough to promote adult education even though this is a better short-term fix for the skills gaps.

The bigger problem with education lies on the science because it takes so long to teach these concepts. But industry has to be aware that it plays an important role too. Unconsciously, companies spend too much time telling people how it does not want more recruits. And those messages soon ripple down to students making decisions that will affect their career choices.

The collapse of the Internet bubble saw electronics engineers sacked from their troubled employers actively discouraging their children from taking up the discipline. That is the last thing an industry needs when it does not have a good image in other respects. If you can't be sexy at least be safe.

In the end, compulsion may be the only answer. A study of the take-up by women of computer-science courses across 21 countries found countries that force children to study maths, science or both almost to the end of their secondary-school careers do better. That does not necessarily extend to more teenagers choosing science-related subjects overall. But, the longer they have the path open to them the better. If people choose arts-only courses from 15 or 16, then they almost certainly will not make it into science when they get to tertiary education.

In the meantime, companies and their representative organisations need to do a lot more to make the science and engineering options look more attractive.

This industry is not for turning
By Chris Edwards

For Malcolm Penn, president of market analyst Future Horizons, the semiconductor industry’s perception of itself has become as troublesome as the more tangible problems of global economic cycles and fab over-capacity. Even though the market grew strongly in 2004, many of those in the industry did not feel as though it was.

The start of 2005 was not great, although it was a long way from the recession that had been feared. Now, the second half of the year is set for robust growth, but there is still a lack of optimism around chipmaking. The problem probably lies with the nature of the recovery that slowly kicked in after 2002 after a slump that rocked the confidence of everyone associated with the industry.

During that time, vendors were happy to make sales and bent over backwards to make sure components orders were shipped in the same quarter. This is no mean feat for an industry where it can take two months to take a raw wafer through to finished product. And those lead times have not improved. Although wafer fab productivity has increased, the newer processes take many more steps. So, it takes as long to make chips as it ever did.

According to Penn, average selling prices have crept up since the slump. But it has been slow progress. And every time the demand flattens out a bit, the need for suppliers to fill their fabs means that prices either stay flat or simply go south. And, all the time, customers are not giving vendors much notice of when they want devices. They still expect turns business.

According to iSuppli analyst Rosemary Farrell, who specialises in watching how inventory moves, customers expect to be able to order turns business, even now that the size of the market exceeds 2000 levels. many of the vendors do not have visibility past the third quarter, even though we are already a third of the way through it. In the last analyst conference call, STMicroelectronics said 90 per cent of its forecast for Q3 was based on orders received so far. And that was in the last week of July, almost one month into the quarter.

The problem for the suppliers is that nothing has happened yet to change the situation. The semiconductor industry has demonstrated that it is able to live with extremely low visibility in orders and has put systems in place to deal with the situation. Those companies who are going to make the most out of this way of doing business are the ones who go further, and decide that actually like it this way, because it is making life hard for their competition. Even though manufacturing lead times are threatening to collide head on with the demand for turns orders, companies are going to have to find ways to manage the problem, or stay unhappy for the foreseeable future.

Unintended consequences
By Geoff Dallimore

18 July 2005

A recent vote in the European Parliament rejected the proposed Directive of the European Parliament and of the Council on the Patentability of Computer-implemented Inventions (CII). In doing so, the European Parliament may have unintentionally supported big businesses over individuals or SMEs.

In its final draft, the Directive would have placed several restrictions on patenting computer programs compared to the current position throughout Europe. More importantly, it would have ensured consistency of approach to the examination and grant of such patents, by patent offices throughout Europe. The European Parliament's decision to throw out the Directive results in the maintenance of a status quo in which there is little legal certainty over the usefulness of patents involving computer programs. A lack of legal certainty usually favours those with more money to spend.

The rejected draft of the Directive was the result of intense lobbying of the European Parliament. At the root of the protests against the CII by various lobby groups appears to be a fear that patents for computer programs might be allowed in Europe - when in fact they already are. Groups such as the Foundation for a Free Information Infrastructure (the FFII) lobbied hard to persuade the European Parliament that patents on computer programs were a significant threat to SME's and that this Directive should be used to prevent any patents involving software from being granted. Whether this was a wilful misrepresentation of the present position in Europe or a genuine ignorance of it remains open to question.

Nevertheless, the open-source community's position was completely at odds with the original purpose of the Directive as suggested by the European Council who felt that limited protection on computer programs was important to the economy, provided that there was clarity and consistency across Europe. The Directive was therefore intended to codify and clarify the current situation, not to substantially alter it. For example, there was never any realistic prospect of the European patent system being opened up to non-technical business methods such as might be patented under current US patent practice. It is therefore no surprise that a massive majority of the European Parliament chose to reject the Directive once the lobbyists succeeded in persuading a few groups in Parliament to introduce amendments that would have severely reduced the level of protection afforded to computer programs compared to the current situation.

The idea for a Directive first came about when it was recognised that differences exist in the protection of computer-implemented inventions offered by the law and practices of different European countries. All European countries now accept that computer programs are patentable under certain circumstances, but the exact set of circumstances varies from country to country. Such uncertainty could create barriers to trade if, for example, a software company was able to obtain or enforce a patent in one European country but not another. So, after wide consultation, the European Commission issued a first draft of the Directive in February 2002, seeking to harmonise this area of law across the EU.

Even in its broadest proposed form, the Directive was never likely to enhance the level of patent protection available to computer programs in Europe. A proposal by the European Council in May of this year was essentially based on the current practice of the European Patent Office (EPO), the largest and most active patent granting body in Europe. The EPO has been granting patents for computer-implemented inventions and the software that runs them for twenty years and there exists a large body of case law regarding the degree to which such patents should be granted.

For example, the EPO will only grant a patent to an invention that makes a "technical contribution" to the state of the art or solves a "technical problem". The European Council's May proposal for the Directive codified this practice of the EPO and made clear that the mere implementation of an otherwise unpatentable method on a computer is not in itself sufficient to warrant a finding that a technical contribution is present. Specifically, the draft Directive stated, a computer-implemented business method in which the only contribution to the state of the art is non-technical cannot constitute a patentable invention. One of the fears expressed by the FFII was that the Directive would make the European patent system as open as the system in the US where patents for business methods and their implementation on computers are routinely granted. These fears were clearly unfounded, but nevertheless seemed to carry a great deal of weight and were one of the reasons for the FFII’s success when lobbying some Parliamentary groups.

Following the Council's proposal, the Legal Affairs Committee of the European Union and others worked on amendments to push the draft Directive towards a more anti-patent position. A Committee session on 20 June formalised a recommendation for a second reading in the European Parliament. Although this was largely misreported in the press as being a shift to a pro-patent position, the proposals actually included some significant anti-software patent elements, such as the requirement for an invention to relate to "controllable forces of nature". Such a requirement might have resulted in technologies which have long been deemed patentable throughout Europe, such as data compression technologies, suddenly becoming unpatentable.

The second reading of the European Parliament took place on 5 and 6 July 2005 and they voted 648 to 14 to reject the Directive. This appears to have followed a realisation by various party groups that an anti-patent Directive would not be allowed by the European Council since it was not in keeping with their original intention.

The European Commission has expressed reluctance to reintroduce the Directive, and the law on patenting computer implemented inventions will continue to develop separately in each European country and at the European Patent Office. The uncertainty in the field of patenting computer-implemented inventions therefore remains and, as this report first suggested, this could be to the advantage of those that have the money to sink into an uncertain patent system. Of course, big businesses as well as SMEs would want clarity since no business likes spending money when there is little certainty of getting something in return. Big businesses were therefore strong supporters of the Directive. Those who chose to oppose it, may have missed a rare opportunity for a more level playing field.

Geoff Dallimore is a chartered and European patent attorney at Boult Wade Tennant.