It began with an ear-splitting roar, futuristic videos of amazing applications replete with multimedia, devices seemingly devised by a tech-drunk engineer, and a lot of big dreams. Third cellular generation, however, has lost a lot of its glamour in the past year. Suddenly 3G didn't seem so imminent. Suddenly it became very expensive. Suddenly it seemed we might not need it at all.
The opening volley of the new era in Israel was fired three weeks ago, when the cellular operators purchased frequencies as a cost of tens of millions of shekels. Market values indicate that herd psychology made companies buy the frequencies, despite the existence of other hush-hush options likely to make the intermediate (2.5) generation more than just a period of transition.
Nevertheless, there are still those ready to spring to the defense of the third generation. Analog Devices Israel Design Center general manager Ran Talmudi, for example, is unmoved by the futuristic vision of high-speed Internet, multimedia, video, and location-based service applications. First of all, he says, there are another one billion Chinese in need of cellular telephones, and it's worthwhile developing another generation just for them.
"Ericsson (Nasdaq: ERICY) claims the downward pressure on prices and the fact that the cellular market is still growing are forcing operators to switch to the third generation," Talmudi explains. "The current and intermediate generation can't supply the demands of the number of users expected in the future. There are 600 million cell phone users today, and China is expected to join the party and bring with it a very respectable number of users from its giant population."
Operation TigerSHARC
Talmudi, aged 51, has managed the Analog Devices (NYSE: ADI) center in Israel, which has 80 employees, since 1996. He began his career in the chip industry in 1979, when he left Elbit Systems (Nasdaq: ESLTF), where he had worked since finishing his studies. He joined the new National Semiconductor (NYSE: NSM) development center in Israel and moved to Davidi Gilo's DSP Group in 1991, which at the time labeled itself as a chip company.
After DSPG decided to split its divisions, Talmudi joined the company employees that founded DSPC, which manufactured cellular communications chips.
Talmudi left DSPC in 1994, five years before its acquisition by Intel (Nasdaq: INTC) for $1.6 billion in what is considered to this day the largest cash acquisition of an Israeli company. For the next two years he worked at nSof (now Digital Appliance Storage Systems), a data storage start-up financed by Oracle (Nasdaq: ORCL) chairman and CEO Larry Ellison.
Following a dispute with the CEO, a group of employees, of whom Talmudi was one, resigned from the company. At more or less the same time, Analog Devices chairman Raymond S. (Ray) Stata was looking for a foothold outside the US. Stata's original idea was to acquire Israeli start-up SmartLink and make it his company's development center in Israel, but SmartLink preferred to remain independent. SmartLink cofounder, president, CTO, and former CEO Benjamin Maytal notified Talmudi of Analog Devices' search, Talmudi submitted his candidacy for setting up and managing the local branch - and won.
The Israeli branch was given the job of developing the TigerSHARC DSP chip. Talmudi says, "It was very unusual for a new branch to be ordered to develop the company's main processor. Even the great Intel tested its Israeli branch for several years before entrusting it with the Pentium. So you could consider it a compliment, but I think it was case of great naivete on the part of both Analog Devices International, which threw us into the project, and the team we set up here, which was confident it could be done."
A penny-pinching mood
Regarding the initial period of adjustment for both sides at the beginning, Talmudi says, ”Analog Devices has been in business for over 35 years. It has been designing analog chips for two-thirds of that time. These chips can be developed by two or three-man teams. A six-man team is considered extremely large. Digital chip design requires more manpower, as well as the use of expensive CAD programs. At the beginning, there were quite a few collisions between us and the US management. We claimed that a development like this needs much more than they were used to, and they thought we had lost our heads. Things dragged a little, and there were delays in spending for the first chip.”
Analog Devices’ penny-pinching mood changed when the chip companies begin preparations for the advanced cellular era. The early talk about cellular third generation and Internet made Analog Devices’ marketing personnel propose that the company enter this field also, and so TigerSHARC was born.
As of the fourth quarter of the company’s accounting year, Analog Devices’ venture into the cellular field has not been particularly successful. In November, the company reported that revenue plunged 88% from $200 million in the previous quarter to $24.2 million and revenue for the whole year was halved to $423 million. Analog Devices explained the fall by weak demand for integrated products for cellular telephones and high-speed communications networks.
The Israeli chip, launched in August 2001, is a DSP chip installed in cellular network transmission stations. The chip designed for high-end performance, (including fixed point and floating point), has a very capacious memory, and supports parallel processing, as well as making it possible to connect a large number of such processors into one unit. The chip bears the internal code name “Nile”, after its older brother “Jordan”, which was developed to demonstrate feasibility and was the Israeli branch’s first product. Jordan was only sold in small quantities, while Nile is designed to be the flagship product, to be sold all around the industry. The next river-named project is the “Danube”, which will contain additional commands lacking in the Nile.
Talmudi says, “The idea is that the chip should handle many calls at the same time. Hundreds of calls take place simultaneously at every transmission station, and every station uses about 100 of our chips. The ability to connect all the chips with each other is a big advantage, and the fact that the chip itself already includes memory eliminates the need to install external memory in the station. Our biggest competitor is Texas Instruments (NYSE: TXN), which also operates with end-user devices themselves and controls most second-generation base stations. But we are making it possible to operate all receiving functions in the base stations using software alone.”
What's so good about the software? Not all the operations required for operating a base station can be done using just the processor – at least, that is how it has been until now. It is therefore necessary to add another chip for those operations. Analog Devices claims it has succeeded in including all these operations in the software, make it unnecessary to use specially-designed ASIC or FPGA components. Saving on components leads to a lower price, which will help any communications provider that has purchased third-generation frequencies for tens or hundreds of millions of dollars (billions, in some countries) and have no money left to finance the stations themselves.
How long will it take before base stations based on your chips are deployed?
Talmudi: ”It will take another 18 months, expect for Japan, where everything works differently, and the network has already been operating for several months. We estimate revenue from the chip developed by the Israeli branch will reach $500 million by the 2006. We won’t have to wait until then to sell – I believe we’ll have substantial sales already by 2003. We have a lot of stamina; despite the recession, we are getting on quite well, because Analog Devices is a diversified company that deals in both the digital and analog fields. At the moment, the analog field is very strong.
”Our goal is 50% of the market. We already have trials in place with top-level companies, such as Ericsson, Nokia (NYSE: NOK), Nortel (NYSE: NT), and Siemens (NYSE: SI; Xetra: SIE). We are fairly new in this field, but are competing rather successfully. In one case, we competed against Motorola (NYSE: MOT), which had multi-year contracts with a multinational company, and swept them aside.”
Employees are pretty secure
Some companies went into both digital and analog and got burned in both of them, because they were unable to organize their activity in both fields. Aren’t you concerned about this scenario?
”Look, Analog Devices makes products with high profit margins, which enables it to survive in the long term. There's a reason we say we deal in both digital and analog – the term that describes us is signal processing, and in this field, we don't move in directions where we don’t want to go. We deal in more exotic fields: conversion from digital to analog and vice versa – fields with much higher profit margins than just digital and analog.”
Does Analog Devices have any other interest in Israel besides the affair with SmartLink which ended on bad terms? Will there be another acquisition?
”Analog Devices has never before bought a company of such size; all its acquisitions have been of small development concerns. There were times, around 18 months or two years ago, when Analog Devices flirted with three companies and I was asked to meet them, but things weren't really concrete, and they eventually gave up the idea. There may be plans to enter new field of which I am unaware. Not long ago, they bought an Australian company dealing in voice to text conversion, and an Austrian company in the XDSL field.”
How much has been invested in the Israeli branch?
”I think $30 million has been invested here since 1996. For a long time, we were a small group of 20 employees; only in the past 30 months have we grown, so for a long time no large investments were necessary. It’s clear we are now feeling a financial pinch, and everyone has to be much more frugal. Once upon a time we flew business class from here to Houston; today we fly business class only if we have to go to Australia. I’m glad to say that we didn’t have to lay off employees in the Israeli branch. Today, we’re even considered the favorites of president and CEO Jerry Fishman, who is concentrating on the cellular market, so we’re fairly safe.”
Ray Stata – the Israeli connections
Even though Fishman, one of the Analog Devices founders, is a Jew, the company’s Israeli connection comes through chairman Ray Stata. Stata was chairman of Libit Signal Processing, an Israeli start-up that was sold to Analog Devices’ major competitor, Texas Instruments, and became that company’s Israeli development center, dealing in cable modems.
Stata held 6.7% of Libit, which held marathon acquisition negotiations with Lucent Technologies (NYSE: LU). Lucent, however, was not quick enough on the draw with its checkbook. Intel also expressed an interest in the company, but in the end Texas Instruments put $365 million on the table for the privilege. According to reports, the talks with Lucent failed because of disagreement over the development center’s location and the fate of the company founders.
Stata was not dismayed, and immediately began negotiations to sell Lucent another company with which he was involved – switches manufacturer Nexabit – for the tidy sum of $900 million. Stata, incidentally, founded Nexabit together with Mukesh Chatter of India, who also cooperated with Stata in two other companies.
Stata also founded his own investment fund, which has already been involved in more than 15 companies, one of which is BandWiz of Israel, which develops broadband content transmission applications. MIT graduate Stata’s link to Israel does not stop there. He also founded OmniGuide, a sort of spinoff based on technology developed at MIT, together with a band of founders that includes Israelis Uri Kolodony and Yoel Fink.
Published by Israel's Business Arena on January 31, 2002