Chips are considered the most turbulent market of all in the technology sector. In addition to the high standards of technology and manpower that are needed in order to operate in the field, the upheaval that the market has experienced, primarily from the demand side, has made the lives of semiconductor managers exceptionally difficult in recent years. In production too, the challenge of having to double the number of transistors every eighteen months (Moore's Law), have increased labor costs considerably. Today, incidentally, there are 560 digital chip designs for lithography of 45 nanometers, and 1,260 such designs for 65 nanometers, out of a total of 10,500 developments registered in 2007.
Israel has seen a good many chip companies. Some of them have been successful, most weren't, but this is a field which is considered to be possibly the last stronghold where the "Israeli genome" has a clear lead over other options. The best example of this, albeit not the most relevant, is the selection by Intel Corporation (Nasdaq: INTC) of the technology developed at its R&D center in Haifa as the technological strategy that will lead its product families in the coming years. The chip field was also responsible for the exceptionally high returns that entrepreneurs and investors earned on activity in the field. Among the companies that "starred" in the history of Israeli technological entrepreneurship, and which were categorized as chip companies, are Galileo Technologies, Passave Technologies Inc., Metalink Ltd. (Nasdaq: MTLK;TASE: MTLK), Mellanox Technologies Ltd. (Nasdaq:MLNX; TASE:MLNX), msystems, Saifun Semiconductors Ltd. (Nasdaq:SFUN), and many others.
"Globes" brought together a panel of participants from almost every possible angle in this industry in Israel: entrepreneurship, investment, research, production, relatively established companies, and tool development for industry. The discussions focused on the changes that the industry has undergone, what is needed from a company in order to successfully compete in the market, the tension between the design and production stages, the feasibility of start-ups be entering the field, financing, and the state of the chip industry in general.
Could you explain, each person from his own angle, what you feel are the big changes in the industry in recent years?
Tower Semiconductor Ltd. CTO Rafi Nave: "I feel that there has been significant change. In the 1990s, the emphasis was on enterprises as end users, and at the beginning of the new millennium there began to be a greater focus on consumer products for the simple user, such as MP3 players, memory devices, cameras, and so on. This has had repercussions in terms of prices, style, extremely tight timetables, as well as the need for mobility, which in turn has triggered a lower rate of power and communications consumption. Another change has been the emergence of China and, I imagine, also that of India in the near future. We will see the consequences of this on both the market and manufacturers."
Metalink Ltd. COO Itzik Ben-Bassat: "Today silicon companies themselves are required to be not just fabless (an R&D center, with production carried out by an outside contractor, S.S.), but also suppliers of silicon systems. Because of pricing pressures, the chain between chip producer and end customer has simply gotten shorter. Today they tell you, 'If you're pitching a product in the cellular or wireless communication field, make sure it comes with architecture at the end product level.' They (the chip producers, principally those in the Far East, S.S.), make the final minor adjustments. This creates pressure in two places: the silicon companies have to add more stages to the product development, while on the other hand, there's pricing pressure because you have to be more efficient, thus gross margins fall. The market size, which has been significantly expanded by consumer electronics, is the factor that compensates for this."
Dr. Yossi Kofman, entrepreneur and former CEO of Modem Art: "One of the trends that is now happening is that chips are becoming more integrative with more components, and we're seeing a lot more System-On-a-Chip items. To put it another way, the need to meet price, supply and location has meant that the threshold for the entry of new players has risen substantially, because of the need to provide a complete system, including hardware, software, and sometimes an application as well, on a single chip. Because of the development over the last decade, most markets are already mature, which means that they are carrying numerous earlier versions that need supporting, so there has been a marked rise in the threshold for new players. One of the outcomes of the maturity that most markets have reached is that they have entered the classic situation where they are dominated by two or three players."
Cadence Design Systems (Israel) Ltd. general manager Gideon Kedem: "I feel that the key issue of the transition to consumer products reflects very strongly. The demands we're now facing are for systems that allow for the design of smaller, cheaper products that include wireless communications, low power consumption, and so on. We need to provide a response to the increasing pressures from the consumer market, while helping companies close the gaps that have been created along the production chain, in order to get the products out on time."
Cadence director business development EMEA region Erez Tsur: "We've seen an 8% fall this year in the number of development activities in the field, and chips are becoming much more complex, a development which has brought about a situation where getting a return on investment in chip development is becoming much harder. We're talking about sales in excess of $500 million in order to justify the development of an advanced chip using 65 nanometer technology. And then people ask you how many markets like this you know of, and how many companies could participate in them?"
Genesis Partners founder and managing partner Dr. Eyal Kishon: "Just as no start-up would compete with Intel or AMD today, so the cellular chip market is maturing, and is now dominated by big companies. An Israeli company seeking to enter the field has to find new things. Another difficulty is that there's no longer any need to reach any given standard, since lots of entities know how to do this. Today, companies have to branch out into new fields that don't presently exist. Cellular, for example, is no longer one of these."
Correct cost analysis is critical
At what stage do you see a bottleneck developing when semiconductor companies make their entry to the market?
Kofman: "The main problem that I see is not at the level of chip development but at the overall system level. If you look at the manpower allocation you will find, to your surprise, that semiconductor companies have the smallest number of VLSI staff (the people that actually design the chips) of all. The big problem is putting together integrative teams with an ability which, essentially, connects the entire system from end to end. Another problem is that in contrast to software, where one can reach revenue pretty quickly and see customers paying, here the time to money is very long and many companies get stuck. At a certain stage, for several years, and sometimes several years from the moment there's a chip, the war is not over money but design wins (the manufacturer's preferred solution, S.S.). This is a tough battle on which a lot of money is spent, and there's no revenue at all."
Ben Bassat: "The financial vulnerability of company like this, until it begins to work under a return-on-investment model, is immense. In software companies, the gross profit margin is almost 100%, and if you hit on the right niche, you can generate revenue pretty quickly. Furthermore, the goal in silicon is to identify the inflection point. You have to guess what will happen in five years. The risks en-route are phenomenal, since standards can be changed by regulatory authorities and you can find yourself betting on the wrong technology. The other risk is identifying the cost models of the end product. You spend tens of millions of dollars on the creation of a product that will sell for a few dollars. You have to guess, based on what you see today, what the models will be in 2010/2011. Such models are difficult to predict. Before you've started working, you work with your suppliers on the pricing models in a few years time. I think that today, correct cost analysis is critical."
Kishon: "The guys here must be obviously be from the silicon field, if they think that it's easier to put together a software company. Launching a product and finding the first customer is easy for them, but the marketing and distribution efforts that come later are very tough and require a lot of money. Each side has its problems. In a silicon company, if you bet on the right product for which there's a real need, your chances of selling a company with a real product but without sales are much higher. So I think the problem here is identifying what the need will be in a few years' time."
IDC Israel research director Gilad Ness: "Standards are one of the variables that really is difficult to predict. Metalink knows this with regard to 802.11n (the next generation of WiFi). Then there's the Ultra Wideband Technology (UWB - a short-range broadband communications technology), for example, which is an amazing instance of how a well-hyped market failed to get off the ground because they spent three years quibbling over the standard, only to find themselves overtaken by something else. In telecommunications technology, the standard can lift a market or kill it. So you either wait for others to push the standard forward and then join them, or you take a gamble and then you could win the jackpot and be the one leading the industry."
Kofman: "One of the problems with small companies, is that in order to gain a foothold in large chip companies, you have to be a certain size. If you want to sell chips to the big firms, you won't even cross the threshold, not because of technological reasons, but because of the supply chain."
Ben Bassat: "There can be no doubting that you need a very strong operational home front, and you need to decide, at the outset, with whom you're going to work. Everyone says TSMC (the largest Taiwanese chip manufacturer), but getting reasonable prices from TSMC is far more difficult. That said, when you come to a supplier and say TSMC, he knows that you have unlimited opportunities, from the point of view of capacity. It does take some time until you see the money, but once you see it, it's a machine that just keeps printing endlessly."
The start-up must precede the market
As regards standards, wouldn't it, perhaps, be better to invest more effort in an aggressive lobby, such as the IEEE?
Ben Bassat: "Metalink invested a lot of money and effort in order to lead the 802.11n standard. Part of the war behind the scenes was to stop the competitors from gaining an advantage, since the big consortiums come along and want to promote their own chips. We teamed up with Intel in order to fight Broadcom, Marvell, and Atheros, and to make sure that none of them wins. Had one of the other companies' pre-standard-implementation (pre-N) designs become the standard, we would have found ourselves out of the market."
Tsur: "One of the factors in the budget is the need to understand the market, and invest in may things, including research, in order to understand the trends. A small start-up can have the technology and the idea, but it also needs breathing space."
Kishon: "This is the point exactly. A start-up must have the technology ready in order to precede the market, otherwise it has no business being there. A start-up can't play the game like Metalink did in this case, with the IEEE. It held two offerings in order to be able to carry this out."
After all these difficulties, how does it feel today to sign a check for a start-up in the semiconductor field?
Kishon: "We've been focusing on problems, but there are also opportunities. We're financing ideas in the field, and there are a quite a few of them. We have cases such as that of Passave, which made an excellent exit with a fantastic growth rate. So you need to identify the people, the companies, and the field. One thing that may, perhaps, have changed is that the amount of money that a company needs today just for development has risen enormously due to the progress in technology."
Kofman: "Nothing has changed in the way you approach investors for money. You still need to come with a vision that looks several years forward, and show things that are sufficiently attractive. It could be that the markets today are slightly different. On the other hand, the capacity for innovation and invention has always been there and always will. On the contrary, chips are becoming building blocks at an ever increasing rate so there are opportunities."
What do you think of Broadcom chairman Henry Samueli's comment that a semiconductor start-up has no chance of becoming the next chip giant?
Nave: "I think that what was true 10 years ago, is still true today. Obviously it will take another 10 years for a company like this to see a billion dollars, but the reality remains unchanged. If the idea is good and the team is good, the sky's the limit."
Ben Bassat: "I agree. The mighty Broadcom too, started out as a division that sold to Linksys, which did some fantastic marketing work. As the barriers are not just technological but also marketing, if you come with a good enough model that's on the inflection point, Broadcom could still lose the market. As a start-up, you don't have to be a Broadcom with sales running into billions, but you need to go in and disturb it. By the time it notices you it will be too late. It's important to do it right, and I think that we could well have a fabless company in Israel with a billion dollar value."
Tsur: "I think there's an element of fear in Samueli's comments. If there's one thing that gives him sleepless nights it's the prospect that a start-up could come along and enter his domain. I think that the big companies are watching start-ups constantly, in order not to miss a thing. There's always a fear that if Intel takes an interest in a certain field, it will assign several hundred engineers to it, but it's not all that simple."
We're very proud of Intel's big R&D center in Israel which, actually, is leading the company, but why haven't any big production centers been set up here? Perhaps the fact that we don't have 10 fabs like Tower is a barrier to the Israeli chip industry?
Nave: "Intel's activity in Israel is one of the most important things that has happened to the Israeli economy. You find Intel graduates everywhere you go, and it wouldn't have been the same without this. Tower's center is different to that of Intel, but it provides a solution to most of the industry's needs. A good many of the ventures have been built on the strength of Tower's presence. We are involved in many of the consortiums initiated by the Chief Scientist, and many ventures begin with us. I think that production is important because it serves as an anchor for industries, but I don't think that we need 10 fabs. We have start-ups which we work with, and it definitely helps them."
Published by Globes [online], Israel business news - www.globes.co.il - on July 26, 2007
© Copyright of Globes Publisher Itonut (1983) Ltd. 2007