Media Eye | How did Marvell achieve a miraculous turnaround in just one year?

A year ago, Marvell encountered a big crisis - the CEO and president both resigned, the company's stock was sluggish, employee morale was low, and Wall Street analysts clamored to sell the company. On June 21, 2016, semiconductor veteran Matthew J. Murphy was appointed as the new CEO and president. He formed a new management team. At the beginning of his tenure, he encountered doubts from the investment field and the industry - can the new CEO lead Marvell to achieve a magical turnaround? In fact, many people did not believe that it could be reversed at that time.

One year has passed, and Marvell has really achieved a miraculous turnaround!! You can take a look at Marvell's stock performance over the past year. Now the company's stock is high, employee morale is high, and new products are constantly being launched. So what did Marvell do this year to achieve a miraculous turnaround?

During the 2017 Taipei Computer Show (Computex), Zhang Guobin, founder and editor-in-chief of Electronic Innovation Network, interviewed Mr. Chris Koopmans, Executive Vice President of Marvell's Network and Connectivity Business Unit. Chris Koopmans just joined Marvell one year ago. As an entrepreneur who once successfully started a business, he interpreted Marvell's miraculous turnaround from the perspective of a "newcomer."

"The first key point for our transformation is focus. Last summer, when our new management team took office, we found that the company had too many different markets in the past. Although we have a great engineering team, we have over-dispersed our investments in some areas that can form an absolute market share. Therefore, our management team reviewed the product portfolio to see which areas are sure to gain a larger market share in the long term. Therefore, we finally focused our core business on three aspects, namely storage, network and connection ." He pointed out, "After finding the key businesses, the next step is to find subcategories for the core business areas, which are areas with greater growth potential in the future. We mainly focus on the enterprise market, data centers and automobiles, which have more growth potential. In addition to fast growth, these subcategories can also help us build our differentiated position, because although some markets are growing very fast, science and technology are developing very fast, which can quickly turn products into channel products ( commodity products ), which makes it impossible for us to use our IP to build field advantages. Our focus helps us build differentiated advantages and make profits. In the past year, our growth in the focus areas has doubled that of our competitors, and our market share has also grown very fast. ( Editor's note: This is actually one of the factors for Marvell's soaring stock price )"

He used the enterprise network market as an example to illustrate the focus strategy. The enterprise network market is the largest segmented market - the fastest growing and only two companies are competing. However, other segmented markets such as automobiles have many competing companies. However, in the enterprise network market, although the demand is strong, it is difficult to increase market share due to the long product design cycle. How does Marvell increase its market share?

He said that the management team hopes to create a culture of innovation and win-win cooperation at Marvell, and hopes to make data-driven decisions in the future to match future growth and the development direction that is compatible with the current IP on hand. "In the past, Marvell's products were only designed based on its own R&D capabilities, but in the future we will develop products that are more targeted at market demand. We will understand the different trends in the enterprise market, the automotive market, or the networking market, and develop our product lines based on such trends. We hope to build an innovative engineering team that can further deepen the relationship with customers and have a keen understanding of the market." He emphasized.

For example, Marvell has found two trends in the enterprise network market. One is that the access layer is developing towards multi-gigabit, while the aggregation layer is developing towards 25GbE. Driven by 802.11ac, wireless connections are developing rapidly. "We have seen that under the new standard, the existing cables of enterprises can only support 1Gb at most and can only support 1G transmission speed. If further upgrades are required, the lines need to be upgraded, which will increase the cost of the enterprise. The new NBASE-T Alliance standard increases the cable rate to 2.5Gbps, so we launched the world's first eight-channel multi-gigabit Ethernet transceiver Alaska® 88E2180 that is compatible with the IEEE 802.3bz standard and the NBASE-T Alliance specification and is designed for CAT5e cable 2.5Gbps and 5Gbps operation.

So enterprises can benefit from Marvell's new PHY technology (data rates can be up to 5 times higher than traditional products). Alaska devices can be used with Marvell's Prestera® switch chip series to provide a comprehensive solution for the transition to a multi-rate Gigabit campus environment. According to him, Marvell's Alaska 88E21xx series is ideal for uplinks on 802.11ax wireless access points, enterprise switches and SFP+ media converters, as well as emerging applications that can utilize Cat5e cabling, such as small cell backhaul, PON residential gateways, high-end PCs and workstations. These devices support the IEEE 802.3az Energy Efficient Ethernet (EEE) standard, which can reduce power usage and energy costs.

The Alaska 88E2180 is uniquely optimized for low power consumption and ease of design, while offering small size and excellent cable performance exceeding IEEE requirements. These features are designed to help enterprise system vendors avoid a series of problems that arise when using other technical solutions, such as expensive investment in development resources, larger PCB boards, and high power budgets in system designs.

This is one product that is part of Marvell's focused strategy, and this highly advanced IC will enable enterprises to quickly and cost-effectively migrate to multi-Gigabit Ethernet network infrastructure.

Why did Marvell discover this business opportunity while competitors did not? Chris said that competitors focus on the high-end market and often use high-end products to attack the low-end market, but Marvell is different. Marvell uses a modular strategy to provide customers with optimized products, which can create more opportunities to grasp the sweet spot of the market. "We have launched 25 new products in this field in the past 18 months. Our customers have made a lot of designs based on these new products, and the new product applications have performed quite well." He emphasized. The following picture shows the products developed by customers based on Marvell.

Many customers watched Marvell's new solution during Computex

In the hot field of autonomous driving, he revealed that Marvell has been able to cooperate with manufacturers in the field of autonomous driving, "because in the field of autonomous driving, there is a strong demand for high-speed access systems. In the past, cars were all point-to-point transmission, and then multi-point to multi-point transmission. On the one hand, data needs to be sent to the driver to see, and on the other hand, it needs to be sent to the autonomous driving processing system. Our network technology is needed, and soon cars will be equipped with similar enterprise-level network products we launched.

At the concurrent forum of Computex 2017, Michael Zimmerman, Vice President and General Manager of Marvell's Networking Business Unit, delivered a keynote speech entitled "Modularization - The Revolution of ODM". In his speech, he specifically pointed out: "British biologist Darwin said that the last organism to survive in the world is not the strongest; nor the most intelligent, but the one with the strongest adaptability." Therefore, the companies that can survive in the current rapidly changing market must be the ones that can best adapt to market changes.

When it comes to modularization, we must talk about Marvell's co-founder Dr. Sehat Sutardja. This talented young man, known as the Indonesian prodigy, went to the United States at the age of 19 and obtained a bachelor of science degree in electrical engineering from Iowa State University. Later, he obtained a master's degree in electrical engineering and a doctorate in computer science from the University of California, Berkeley. Dr. Sehat Sutardja proposed two innovative technologies in 2015 - one is the final level cache (FLC) technology, and the other is the MoChi modular chip technology.

Mochi is to use a new interconnect technology (Mochi) to realize the functions of SoC, reduce R&D and production costs, and speed up time to market. The MoChi interconnect chip is based on the ARM AXI link running at up to 8Gbps or even faster, which can maintain a very low chip-to-chip latency. The MoChi link can connect multiple chips together in a daisy chain, and can realize compact serial/deserializers (micro-serdes) and low-voltage differential signals.

The core of this MoChi solution is to use a high-speed, low-latency in-house SerDes interface to quickly package existing dies together using TSV technology according to needs.

Chris said that Marvell will use this architecture to build a product portfolio in the future and will further develop more product lines.

Chris said Marvell is a company driven by an engineering culture. The company currently has 9,000 technology patents. "In the future, Marvell's innovation will focus on areas related to high-speed data transmission, such as analog-to-digital conversion and other high-speed information transmission technologies. We want to build a platform based on high-speed transmission technology for our customers to use. This is a principle we uphold." He pointed out.

In terms of supporting the Chinese market, he said that the Chinese market is growing very rapidly and plays three important strategic support points for Marvell. The first is the source of R&D talents. China is Marvell's R&D center and one of its three major R&D centers. Marvell will continue to recruit talents in China. Second, China is an important base for Marvell's supply chain partners. Marvell has many long-term partners in the Chinese market. Third, China is Marvell's terminal product sales market. The Chinese market is also an important market for Marvell's enterprise category, data center, consumer, and automotive segments. The Chinese market is growing faster than other markets. "We attach great importance to this market, as well as our partners and customers. We are very satisfied with the Chinese market." He pointed out.

In fact, strategic focus is a commonplace topic that many corporate decision makers understand. However, when you really have to make a bold decision to cut off a production line that brings you revenue, it is a test of the manager's courage and insight. I hope Marvell's miraculous turnaround can bring inspiration to local IC companies.

Note: The article is reproduced from WeChat public account Zhang Guobin