It is now just over 19 years since the day - at least so the legend in the chip industry has it - that Dr. Sehat Sutardja, and his brother Dr. Pantas Sutardja attended a meeting with one of the leading scientists at Seagate Technology (NYSE: STX). What was said at that meeting is of little importance. But the outcome changed, perhaps for the last time in history, the balance of power in the industry. Two years later, Marvell Technology Group (Nasdaq: MRVL), the company that the Sutardja brothers and Sehat's wife Weili Dai founded, became the largest supplier of chips to the hard drive industry, deposing Texas Instruments Inc. (NYSE: TXN), and then NEC Electronics Corp. (TSE: 6723), and Infineon Technologies AG (NYSE; XETRA: IFX).
A good many electrons have passed through printed circuits since then. Marvell has gone through a lot too, expanding in several directions and entering new markets. It also managed at the same time to swallow a number of technology companies with considerable success, and develop several more.
The basis on which the company was founded - the Sutardja brothers' belief that they could produce chips better than anyone else - has remained consistent. But the execution has been a different story. It turns out that the forces at play in a company with 5,500 employees and a market cap of $6.5 billion are very different when it attempts to move as fast as it was used to.
Marvell's captains are now discovering the mathematics themselves. Over the past year, since it announced the acquisition of the Intel wireless processor product line (the XScale division) for $600 million, the company's share price has fallen 50%. The situation failed to improve even after the company announced it would be cutting expenses and jobs. On the contrary, in the six weeks that have passed since the announcement, the stock has fallen a further 35%. "We believe we have the technology and the people," says Suhat Sutardja in an exclusive interview with "Globes" during his visit to Israel this week, "and ten years from now very few companies will remain as large suppliers, following the consolidation in the chip market. We want to be one of those that survive," he declares.
Globes: You announced a 7% cut in your workforce and the stock fell sharply in response. Perhaps the investors were expecting a more aggressive cut?
Sutardja: "Fortunately, that's not the feedback we've been getting from investors. They're more concerned about the limited information from the company. We haven't had a CFO to liaise with them for the past 6-7 months (the previous CFO, George Hervey, resigned in May following an inquiry by the US Securities and Exchange Commission into options backdating at the company, S.S). We're not as good as we once were when it comes to communicating with investors, but we hope to solve the problem. The investors understand that we did what we had to do."
At least some analysts feel the cut should have more drastic.
"Very few investors have said that to me. Their main concern is that we should get on with our jobs. As company CEO, it is my responsibility to ensure we don't make too steep a cut in our workforce, because if we do that we won't manage to develop the things we need on time. It may improve the bottom line quicker, but it will be felt over the longer-term. The lay-offs we made were in the proper balance. Anything more would be too much and impractical."
And on the other hand, is there a chance that you will expand your activity in the near future, through acquisitions or recruitment drives?
"The key thing is that as a company we have to find the balance between the speed at which we can get things done, and profitability. We had to move really fast over the past seven years in terms of acquisitions and workforce expansion in order to make a quick entry to the various communications markets, and our profitability suffered as a result of this, so we had take a difficult step a few weeks ago, to ensure our profitability improves. We will continue to develop new technologies, but we need to be careful not to expand too quickly until we're strong enough."
From Galileo to XScale
Marvell began its rapid growth trail with the acquisition of Galileo Technology Ltd. in 2001. The acquisition, which was made for $2.7 billion in shares (and subsequently shrunk by two thirds by the time the deal closed) was considered a highly successful gamble and within a short time it made Marvell one of the leaders of the communications chip market.
Galileo developed switched Ethernet controllers and, according to Sutardja, it was successfully integrated into the infrastructure that Marvell had at the time. "Information processing, storage, transmission, and display - these are the four key things that you need to address if you want to supply technology to the consumer market. The communications element was a perfect fit with what Galileo provided us," he says.
Several acquisitions on, the success with Galileo appears to have been difficult to repeat. Marvell never received any credit for its execution. It knew how to extract the best out of every acquisition it made, but the XScale division, while it is not the largest or the most significant of Marvell's acquisitions, has not yet delivered the results the company was hoping for. "We look at the Galileo deal seven years later, and all we see is success," says Sutardja. "Success is the only thing people see, but integrating a large number of people into a new company is always a challenge, and there were difficulties then as well. It should be remembered that when we integrated Galileo, it was still a relatively small company."
At any rate, for Marvell it was the first time it had acquired a loss-making business managed in a manner diametrically opposed to its mentality. Intel, a large organization with a different modus operandi, did not manage to put the business on a track suited to the cellular processor markets, where the margins and prices differ radically from the markets it operates in traditionally.
"In the Galileo acquisition we were about 350-400 people on each side," Sutardja says, by way of comparison. "So relatively speaking, it was easier (the Intel division had 1,200 employees when Marvell acquired it, while Marvell itself employed 4,500, S.S.). Integrating this number of people is more complex, but the recipe for success is similar - you have to spread the appropriate culture across the company."
Marvell ended the third quarter of 2007 with sales of $760 million, 45% more than in the corresponding quarter of 2006, and a loss on a GAAP basis. Most of the growth was based on $140 million in sales of products for the cellular market and palm top computers, following the acquisition of the Intel division. The sales in this business were high (and are expected to total $500 million for 2007 as a whole), but they generated an operating loss of 21%.
The good news is that the losses are narrowing. Marvell ended the quarter with an operating loss of 19% of the unit's sales, compared with 40% in the first quarter of the year. Analysts at Citigroup believe the unit will begin showing an operating profit by 2010.
Can you estimate when this activity will become profitable?
"One of the big challenges was to move production from Intel's fabs to the ones we usually use (such as Taiwan Semiconductor Corporation, S.S.). Last week we announced that we had completed the move and this is a big step. The next step on the way to profitability is to design a new silicon wafer for the chips, using our production facility infrastructure. As far as the timeline is concerned, both these steps should be completed within the next one to two years."
Chips are twice as intricate
The world that Marvell operates in is based on advanced technologies and tremendous costs, thereby requiring it, as a fabless company, to move at lighting speed. This speed is required mainly in order to grab market share. As a consequence, both Marvell and its competitors feel an increasing need to make acquisitions, a need driven by falling growth rates, as the figures, in Marvell's case at least, show.
The cause, as mentioned earlier, is the market Marvell operates in. Its traditional product was a processor for external storage - the hard drive. The company now controls a 40% share of the market and, in the last quarter at least, this business accounted for 40% of its revenue. If we look at the overall figures, there has been virtually no growth in sales in the storage chip segment. Nor has there been any tangible growth in sales of the company's enterprise communication product line (Ethernet products). Marvell's sales in these two segments accounted for 60% of total sales in the last quarter.
"We're trying to target major opportunities in chips principally in the consumer sector, but also the enterprise sector," says Sutardja. "The technology industry is constantly moving forward, so we will eventually need additional technology. If you look at what we acquired from Galileo, it would have taken us ages to develop it ourselves. Theoretically, we had the ability to develop it in-house, but in the meantime the market would have gone elsewhere."
Given these pressures, can a start-up succeed in this field today?
"Theoretically, yes. The only difference is that 12 years ago, each chip was a lot less complicated. The level of complexity doubles every year and half or so. So in the last 12 years, the complexity has increased twenty fold. Therefore, any company which wants to go into the business today has to develop a lot of components in order to start off at the place where the market is. So theoretically it's possible, but it's much more challenging."
Marvell faces a challenge, not just in gaining market share, but also from the industry itself. It reported larger-than-expected demand in the last quarter, but the analysts were still not impressed. Friedman, Billings, Ramsey, for example, recently wrote that the chip business cycle could face a downturn in the near future.
There are fears that the chip business is on the verge of a slowdown. How do you see it?
"I'm not all that worried. I am more worried about the prospect of a slowdown in the economy. Rising energy prices and the crisis in the mortgage market, these are two issues that I should be worried about, since they can have an impact on consumer spending. If that happens it will ultimately affect the chip market too. Should I be worried about a global recession? Yes I should."
"We'll invest $200 million in Israel a year"
Marvell first began operating in Israel in 2001 with the acquisition of Galileo. The company felt that the aggressive way Israelis do business suited its own management style and set up its second largest R&D center in Israel. The Galileo deal was Marvell's first major acquisition and its first in Israel. The Israeli operation was named Marvell Semiconductor Israel and managed from Yokne'am. In 2003, Marvell made a further key move locally when it acquired Radlan Computer Communications Ltd., which focused on embedded networking software, and following the acquisition of the XScale division from Intel in 2006, it took over Intel's R&D center (which it inherited from DSPC) in Petah Tikva.
"When we bought Galileo we were younger than it was, and it took a lot of courage from us to carry the deal through," says Weili Dai. "We went on to acquire Radlan, and in 2006 we acquired DSPC. We have the largest fabless center in Israel and we will expand it even further."
Marvell Semiconductor Israel is managed by Eliaz Lavi, who was appointed overall manager of all the company's operations in Israel six weeks ago. After seven years of making the news over its acquisitions in Israel, Marvell is now in the unenviable position of making the headlines because of its job cuts, and the company is now adopting a new approach to its relations with both media and its staff.
Marvell fired 106 people in Israel, most of them at the former Intel division, and according to Lavi, there will be no more changes after this, and that the organization is now being stabilized. "When you see Marvell's rapid growth, it's very natural to look back at what you've been doing after a while, and try to optimize the activity," says Lavi. "If you look at Israel, we had four management teams, and now we have one. I think that a company that does not take these steps is destined to fail."
What are the plans for the future?
Lavi: "Marvell has invested $3 billion here. We have the people, the technology, and the vision. So what will happen? We'll grow, but it will be done gradually and carefully. We won't rush out and hire people like lunatics, because you can't hire people and then fire them immediately afterward."
Marvell spends $200 million a year on its activity in Israel. Sutardja says the company intends to invest $600 million in Israel over the next three years, of which 60-70% will be salary costs, and adds that the company intends to increase its expenditure in Israel by a further $40 million during the same period, which will be spent on what he calls "infrastructure."
An extract from this interview was published by Globes [online] on January 14, 2008.
Published by Globes [online], Israel business news - www.globes.co.il - on January 17, 2008
© Copyright of Globes Publisher Itonut (1983) Ltd. 2008