4401 South Beltwood Parkway
Dallas Semiconductor Corporation designs, manufactures, and markets electronic chips and chip-based subsystems. The Company continually invests in multiple, non-commodity products using advanced technologies to reach diverse markets and customers. Founded in 1984, Dallas Semiconductor's strategy has resulted in diversified revenues, balanced end-markets, and a global customer base.
Dallas Semiconductor Corporation designs, manufactures, and markets a wide variety of semiconductors and semiconductor-based subsystems used in computers and other electronic equipment. Its product mix is one of the most diversified in the market, with some 350 proprietary base products and more than 2,000 variations. The company sells its components directly to manufacturers of personal computers, industrial controls, automatic identification devices, telecommunications equipment, and scientific and medical equipment, among others. With more than 15,000 customers worldwide--including Compaq, Hewlett Packard, IBM, Intel, Lucent, Motorola, Nortel, and Sony--Dallas Semiconductor enjoys one of the more balanced customer mixes in the semiconductor industry. These factors have helped it successfully navigate the volatile semiconductor market, posting record sales every year until 1998, when industry conditions caused a drop in sales. That year sales totaled $342.6 million, with a net income of $55.4 million.
The continued success of Dallas Semiconductor is perhaps the product of a lesson well learned. Dallas Semiconductor was established in 1984 by C. Vincent Prothro (chairman, chief executive officer, and president) and several of his former colleagues from the Mostek Corp. From 1977 to 1982, Prothro had served as president and then chief executive of Mostek, a subsidiary of United Technologies, a company that pioneered the development of random access memory chips. In 1980 Mostek was an incredibly profitable firm. It was one of the few manufacturers of a random access memory chip, known as a DRAM, that had quickly become a standard component in personal computers and other electronic products. In 1981 Mostek enjoyed 55 percent of the world market and $360 million in sales. The following year, however, sales plummeted by $160 million to $200 million. One year later, sales were even more dismal.
During the early 1980s, a number of competitors in Asia and the United States had developed less expensive ways of making a chip capable of performing the same function as Mostek's. 'With almost no warning, we went from having a wonderful time to being in great jeopardy,' Prothro told Investor's Business Daily in 1993. The semiconductor war was on, and the market became treacherous. One of Mostek's problems was that it hadn't diversified either its product line or its customer base: 70 percent of its sales were of the same product, split between five customers.
In 1983 Prothro left Mostek and with a number of partners formed Southwest Enterprise Associates, a high technology venture capital fund. The following year, Dallas Semiconductor was established as Prothro and Southwest Enterprise joined forces with Dr. Chao C. Mai and Michael L. Bolan, former Mostek executives, and John W. Smith, Jr., a venture capitalist who assumed the duties of president, chief executive officer, and chief operating officer.
Product and Market Diversification, 1984--87
The founders recognized the need for a strategy that would prevent the company from relying on a single market, product, customer, or technology. 'The semiconductor industry long ago learned to make the same product cheaply,' Bolan told Fortune in 1985. 'Our goal is to make unique products cheaply.' Its strategy was to address specific client needs, develop a solution, and then later adapt the solution to fit other client needs.
One of the primary ways Dallas Semiconductor sought to do this was to allow clients to customize chips at a very late stage in the production process. Known as 'late definition technology,' the process used ion implants, lasers, or embedded lithium to etch chips according to precise customer specifications. This process gave Dallas Semiconductor a great deal of flexibility in adapting to changes in the marketplace and also provided the foundation for the company's versatile product mix. 'All the major developments in the industry have been decided by the marketplace, not by planning,' Bolan said in 1985, adding that 'Our cardinal promise is not to fall in love with any one product.'
Another innovation was a product whose origins could be traced to Mostek: the pairing of lithium batteries--a small, long-lasting battery found in wristwatches--with a random access memory chip called a complementary metal oxide semiconductor or CMOS. CMOS chips were used to power individual software programs and were popular because data could be added to or erased from them at will. Their main drawback, however, was that whenever the power supply was cut from the chip, all stored data was erased from the chip's memory. With this new product, known as nonvolatile RAM, the lithium battery would continue to supply power when electricity failed, allowing the chip to be powered continuously for over ten years. The development was essential for lifesaving medical equipment and for cash registers or banking systems that keep running totals of deposits and other transactions.
By March 1985 Dallas Semiconductor had regional sales offices in Philadelphia, Indianapolis, and Phoenix, and had begun filling orders for a series of nonvolatile RAM it had developed based on the lithium battery technology. The company also introduced a line of electronic chips based on the same technology that prevented software theft. The product consisted of a key containing an integrated circuit coupled with a lithium battery and a corresponding socket inserted into the computer. Software manufacturers could encode the key's circuit to match a code in the software package, and only when the user inserted the key in the socket could he or she use the program. Although theft was one of the primary concerns of software manufacturers, this product was not as successful as the nonvolatile RAM, most likely because software buyers were reluctant to purchase and install the socket.
In its first year of operation, Dallas Semiconductor posted sales of approximately $3 million, slightly below its projected figure of $3.5 million. In 1986 the company received $10.8 million in a second round of capitalization, a move that increased its total capitalization to $32.2 million. Construction continued on the company's fabrication plant, which had already grown to 65,000 square feet. By late 1986 the company had introduced a line of application-specific telecommunications products, primarily T1 circuits capable of supporting a high volume of voice and data transmission.
Continuous Growth as a Public Company, 1987--97
In 1987 the three-year-old Dallas Semiconductor went public on the New York Stock Exchange at $9 per share. The initial public offering brought in an estimated $30 million, which was funneled into research and development. This led to the creation of 'one conductor semiconductor technology' which, when coupled with the company's other technologies, formed the basis of Dallas Semiconductor's automatic identification systems and Touch Memory chips. In 1987 the company also introduced a line of highly successful computer clocks, a simple self-powered component added to computers that could keep time for ten years, regardless of whether the machine was turned on. One version was even capable of 'waking a computer up' to perform a specific task and then shutting it down after the task was completed. Dallas Semiconductor quickly became the market leader in computer clocks, a position it continued to hold throughout the 1990s.
In its first four years of business, Dallas Semiconductor's annual sales increased almost tenfold to $30.6 million, and net income grew to $2.6 million. Sales jumped in 1989 to $58 million, while net income grew to $9.9 million. All income was fueled back into operations under the conservative directorship of Prothro, who had begun assuming many of the executive duties held by Smith. In 1989 Smith retired and Prothro assumed position of chairman, chief executive officer, and president of the company.
In 1991 Dallas Semiconductor borrowed a concept from Post-It Notes and introduced an innovative line of data storage and retrieval systems called Touch Memory. The system consisted of a memory chip stored in a stainless steel case about the size of a nickel. On the back of the case was a band of adhesive that allowed the chip to be attached to anything from goods moving through a factory to employee identification tags. Data could be stored in and retrieved from the chip using a simple metal probe linked to a handheld computing device, personal computer, or factory controller. Because the chips could be updated at will, they were more flexible than bar codes or printed tags; because they were small, they could be placed just about anywhere.
Dallas Semiconductor soon began customizing this system for a variety of different applications. In 1993 the U.S. Postal Service purchased the system to monitor drop box mail collection and to provide carriers with an efficient means of reporting on the condition of the postal box. Ryder Systems Inc., operator of a nationwide truck rental service, purchased more than 1,000 systems of Touch Memory in 1994, which it attached to rental trucks and used to store maintenance records. Another variation of the system was used to replace the punch cards that farm workers used to count the number of bushels each farmer picked. This system allowed farm owners to track each worker's daily harvest more efficiently and simplified payroll procedures. However, according to the Wall Street Journal, it also caused concern among farm workers unions, due to the possibility that the 'system would allow farm owners to track workers from job to job, in effect keeping a record of their employment history.'
By 1993 Dallas Semiconductor had developed more than 170 base products in 14 different categories. Its core products included telecommunications and timekeeping systems, nonvolatile RAM, automatic identification systems, and microcontrollers. 'Individually none of the products are showstoppers,' an analyst told Investor's Business Daily that year, 'but collectively they provide nice growth and better than average earnings.' Net sales totaled $156.8 million to more than 8,000 customers. For 75 percent of those customers, Dallas Semiconductor was the product's sole supplier.
The company continued to pour its earnings into research and development, creating novel applications for existing technology and building a new wafer fabrication facility that doubled the company's production capacity in 1994. In 1993 Dallas Semiconductor was nominated for EDN Magazine's 'Innovation of the Year' award for a high-speed microcontroller that was three times faster than existing micros. Touch Memory continued to show promise as the company introduced an Emissions Memory Tag, which allowed chemical companies to monitor emissions for compliance with the 1994 Clean Air Act. The company also successfully transferred its digital technology to create the first digital thermometer that could be read directly by a computer. And in 1994 the company introduced a microchip said to prevent even the best computer hacker from entering corporate systems. Called Dallas SignOn, the chip used the same technology as hotel security cards and was allegedly 100 percent tamperproof. In 1994 overall sales grew for the fifth consecutive year, to $181.4 million, although sales in some core products began to slow as their markets matured.
In just ten years, Dallas Semiconductor's sales had grown from $3 million to more than $180 million. In the mid-1990s, the company made the Forbes list of the 200 best small companies in America, and its future seemed healthy. This was largely attributable to Prothro, known as a 'cost-conscious manager who is focused on the bottom line and gets by with moderate resources.' Prothro's management style was perhaps best characterized by the company's efficient use of existing technology and its steady, measured growth. The company had $75 million in reserve that would serve it well if the semiconductor market were to take a sudden turn for the worse. But more importantly, it had a diversified line of products which it sold to an equally diverse customer base in a variety of markets.
As further proof of its growing financial strength, Dallas Semiconductor declared its first dividend ($0.025 per share) in the first quarter of 1995. Future plans included efforts to expand its markets in Europe and Asia, and to increase growth through acquisition. In 1994 the company proposed a merger with RF Monolithics Inc., a manufacturer of remote and wireless technologies. However, RFM was in the process of going public, and the deal was never consummated. Most likely this was due to Prothro's concern with the bottom line. 'When we put this company together, we wanted it to be self-funding, not a giant consumer of cash,' Prothro told Investor's Business Daily in 1994, adding, 'We'll take the slower growth path if we can become more profitable as a result.'
Annual revenues continued to grow, reaching $233 million for 1995, $288 million for 1996, and $368.2 million in 1997 before falling back in 1998 to $342.6 million, a drop largely attributable to weak market conditions. It was the first drop in sales revenue in the company's history. During this period the company's diverse product mix and avoidance of commodity products gave it a stability unusual for the semiconductor industry. Its stock did not rise as fast as the overall industry, but when there was a downturn, it did not lose as much value, either. In May 1997 insiders at the company sold between 44 percent and 86 percent of their actionable holdings following a surge in the stock price after a strong first quarter. At the beginning of 1998 there was more significant insider selling, despite analysts' predictions for a strong 1998. In September 1999 the company adopted a Shareholders' Rights Plan to discourage a hostile takeover; the plan would go into effect whenever a person or group acquired 15 percent or more of the firm's common stock or announced a tender offer to do so.
About half of the company's production was exported outside the United States, with 30 percent going to Asia and 20 percent to Europe. Although it had a presence in 40 countries, all manufacturing was done in Dallas. The company invested a record $60.5 million in 1996 in added fabrication and test equipment, facilities, and design automation tools. Its facilities took up 40 acres.
It was around this time that Dallas Semiconductor began developing its line of iButton branded automatic information products. These were products that could be attached to an object or carried by a person and that identified the user and held relevant information. In 1996 Dallas Semiconductor came out with a wearable security chip called the Cryptographic iButton, which was expected to provide a solution to ensuring the identity of Internet users, especially mobile ones. The product included a processor, true time clock, random number generator, arithmetic accelerator, and 8KB of SRAM. The first generation of the product was originally introduced in 1992 as a memory device that electronically stored cash and was used in Turkey and other countries to pay fares on the transit system. The new product--a "computer on a chip"--functioned as a security device by positively authenticating an individual to a server-computer or to another person through possession of the button and a personal identification number (PIN). One of its first applications was through the U.S. Postal Service (USPS), which distributed it to customers using postage meters. By 1999 Dallas Semiconductor had shipped more than 30 million portable data carriers under its iButton brand for a wide range of applications, from the labeling of printer cartridges to physical access control keys for Federal Reserve Bank vaults.
Dallas Semiconductor continued to work with the USPS, and in 1998 the company provided a postal-security device for the first Internet electronic-stamp delivery system, which was developed by E-Stamp Corporation. Beta testing of the system was approved by the U.S. Postal Service on March 31, 1998. The system allowed users to purchase and download electronic stamps to an electronic postal-security device attached to their computer's parallel port. The device protected and kept track of remaining postage. E-Stamp provided Windows-based software to interact with the device and print the stamps.
The postal-security device was a new version of the Crypto iButton, with an 8051-compatabile multiprocessor, a real-time clock, 32KB of ROM, 6KB of nonvolatile RAM, and an exponential accelerator for integers up to 1024 bytes. The electronics were packaged in a heavy-duty, stainless-steel housing with tamper-detection circuitry that immediately erased critical data when there was an intrusion.
In August 1999 the USPS and Dallas Semiconductor jointly announced that the cryptographic iButton had successfully completed laboratory testing under the direction of the National Institute of Standards and Technology (NIST). The testing was a prerequisite for the iButton's role in an Internet-based system for issuing electronic postage on home computers and printers. The USPS referred to these electronic stamps as Information Based Indicia and called the iButton a Postal Security Device (PSD). Dallas Semiconductor planned to deliver an adapter for use with Microsoft's release of Windows 2000 as well as a Java-powered version.
Also in 1999 Dallas Semiconductor introduced a variation on the iButton called the Thermochron iButton, a reusable time- and temperature-logging device. It could monitor data in two ways concurrently, one by recording the temperature and date, and the other by recording the frequency of thermal events.
Commitment to Conservative Growth: 1999
Another manufacturing expansion took place in early 1999 when the company opened a new automated manufacturing facility to produce its proprietary 1-Wire chips in packages no bigger than the chip itself. The chip employed a unique solder bump that enabled it to merge bi-directional digital communication and power into one signal plus a ground return. The minimalist chip was designed to save space in portable and other space-constrained applications.
Dallas Semiconductor's strategy of investing in non-commodity products for diverse markets and customers has served it well. Its largest customer accounts for less than five percent of the company's total sales, and its top 25 customers account for less than 30 percent of total sales. The firm's 350 base products and their variations are grouped into seven major product categories, giving the company a basis for stable growth that is unusual in the semiconductor industry. After sales declined to a low point in the third quarter of 1998, they have risen steadily for three consecutive quarters. As the company prepared to announce results for the third quarter of 1999, its stock was trading at a 52-week high.
Principal Competitors: Analog Devices; National Semiconductor; Rockwell International; Microchip Technology; STMicroelectronics; Texas Instruments.