Incorporated: 1998
NAIC: 541512 Computer Systems Design Services
With its headquarters located in Cambridge, Massachusetts, Akamai Technologies, Inc. maintains a massive distributed computer network, composed of more than 15,000 servers in more than 65 countries. The network provides software and services to help companies, government entities, and other customers ensure reliable delivery of web content and applications. Rather than just supplementing a customer's server capacity, in effect acting as a mirror site or caching service, Akamai uses sophisticated programming to monitor web traffic, makes extra copies of high-demand content, and spreads out requests, routing them to a server in the geographic proximity of the end user. In this way, customers do not have to maintain an excessive number of servers to handle peak traffic and are able to weather usage spikes without overloading their systems. In addition, Akamai offers an array of professional services, including Internet architectural design, implementation, performance optimization, and security. Akamai maintains offices in nine U.S. cities, as well as offices in Europe and the Pacific Rim.
Originally developed by the U.S. military, the Internet was little more than a collection of linked computers before Dr. Tim Berners-Lee, a British computer scientist, invented a software program in 1989 called the World Wide Web, which provided easy access to the quickly accumulating resources of the Internet. The program supplied an address (universal resource locator, or URL) to documents, which were encoded with hypertext markup language (HTML), and linked for easy retrieval by hypertext transfer protocol (HTTP). Coupled with a graphic interface, Mosaic, which became the backbone of web browsers, the Internet enjoyed explosive growth in the 1990s. Berners-Lee joined the faculty of the Massachusetts Institute of Technology (MIT) in 1994, heading the W3 Consortium for the Laboratory for Computer Science, very much devoted to making sure the Web developed into the universal tool he envisioned it to be. But by 1995 he became concerned that traffic problems could have a deleterious effect on the growth of the Web, in particular the "hot spot" or "flash crowd" phenomenon, when large numbers of web users try to access the same site at the same time, leading to a bottleneck that would not only bring down a site but create a domino effect, with the network around the site becoming a casualty. Moreover, the future of the Internet was jeopardized because of a general lack of speed, leading wags to call the Internet the "World Wide Wait." In 1995 Berners-Lee issued a challenge to his MIT colleagues to improve the way in which Internet content was delivered.
Among the people who accepted the challenge was Tom Leighton, a professor of Applied Mathematics, who headed the Algorithms Group at MIT's Laboratory for Computer Science. Although initially regarded as an academic exercise, the task of speeding up the Internet offered MIT math students a chance to work on a problem with a real-world application. Over the next two years, Leighton, his colleagues, and graduate students worked with little sense of urgency, since no one at the time gave any thought to building a company around the results of their research. Among the graduate students who used this work as a basis for their theses was Daniel Lewin, who devised key algorithms that would become part of a system that efficiently routed Internet traffic, achieved in part by spreading copies of high-demand material over a network of distributed servers.
A chance conversation with an MIT business student led Lewin to suggest that he and Leighton use their research to enter MIT's Sloan School of Management's annual entrepreneur contest. Lewin's motivation was simple enough: He needed his share of the $50,000 prize to pay off some student loans. Although their business proposal failed to win the 1998 competition, it was among the six finalists out of 100 entries, prompting Leighton and Lewin to continue their work. By August 1998 they had developed a prototype of their system that worked to their satisfaction. Some major customers interested in working with them on beta trials were lined up, and the partners began taking steps, including the licensing of some MIT intellectual property, to form a company to exploit their research. For a name they chose Akamai, a word they found after coming across a Hawaiian dictionary on the Web and looking up the word for "intelligent." Akamai was also a colloquial term for "cool."
Akamai found no shortage of interested investors. In a first round of venture capital, the start-up company received $8.3 million from Boston's Polaris Ventures and New York's Baker Communications. A second round was conducted a few months later, garnering $35 million from Polaris and Baker, plus newcomers Battery Ventures of Wellesley, Massachusetts, and TCW Group of Los Angeles. In addition, Akamai was able to sell interests in the company to other high-tech companies. Akamai and Cisco Systems Inc. began discussing a technology-sharing agreement, which led to Cisco buying a 4 percent stake for $50 million, followed a few weeks later with Microsoft Corporation acquiring a 1 percent stake for $15 million. To that point, Akamai's software was being run on a Unix-based operating system, Linus. The Microsoft deal called for Akamai to develop a version of its software that could be run on Microsoft's NT operating system. Later, Apple also paid $12.5 million to get a piece of the action.
Akamai's business model also attracted the attention of seasoned executive George H. Conrades, a 30-year veteran of IBM who rose to the rank of senior vice-president for U.S. Operations and later headed technology research and development firm BBN until it was acquired by GTE Corp. He then became a partner at Polaris and after the venture capital firm began investing in Akamai he became a member of the start-up's board and served in an advisory capacity. He was subsequently tabbed to serve as Akamai's chairman and CEO, taking over in April 1999 when the company became operational. Akamai's chief executive officer, Paul Sagan, was another experienced hand, the former president of Time Inc. New Media and founder of Time Warner Cable's Road Runner high-speed cable service.
From the outset Akamai boasted an impressive roster of clients, including Yahoo, CNN, About.com, and Go Network. The first product was called FreeFlow. The service, as described by Network World in a January 1999 profile, "delivers customers' web pages via Akamai's global network of distributed web servers. ... The Akamai network determines where hits are coming from and shifts copies of the pages sought to Akamai servers nearest the source of the demand. When demand drops, the network cuts back on the number of servers delivering content. Customers don't have to add hardware or change Internet access to use the service. With faster response times, Akamai customers will be able to post more complex pages." FreeFlow was able to achieve this feat because "the network uses server software based on a blend of four families of algorithms: randomized, online, flow and consistent hashing. The software runs on each server, distributing intelligence around the network so adjustments are made without intervention from a central site. ... If one server goes down, others become aware and pick up the slack." FreeFlow was put to the test in October 1999 when it was used by NetAid to support the webcast of a concert to draw attention to the problem of world hunger. Despite the immense traffic to the NetAid site, there were no crashes or delays, a far cry from the debacle that took place several months earlier when Victoria's Secret attempted to webcast a fashion show without the support of a dynamic system like FreeFlow.
Very quickly Akamai became a Wall Street darling, involved in a sector (essentially any company even remotely connected to the Internet) that was proving irresistible to investors. The company took advantage of the interest by announcing in August 1999 an initial public offering (IPO) of stock, in the hope of raising about $86 million by selling shares in the $16 to $18 range. Morgan Stanley Dean Witter served as lead manager of the offering, joined by co-managers Donaldson, Lufkin & Jenrette, Salomon Smith Barney, and Thomas Weisel Partners LLC. But when the IPO was actually conducted in October, demand for shares was so strong--due in no small measure to the success of the NetAid concert--that Akamai commanded a price of $26 a share, selling 9 million shares to raise $234 million and netting $216.5 million. Moreover, as soon as shares began trading on the NASDAQ they soared in value to more than $145 a share, a 458 percent gain in the first day on heavy trading, the fourth steepest climb ever for a U.S. stock in its first day. As a result, 42-year-old Leighton and 29-year-old Lewin were worth on paper $1.4 billion each. Conrades owned enough stock to be worth nearly $960 million. Even MIT graduate students who performed summer and part-time work for stock options were now millionaires. Akamai itself had a market capitalization of about $13 billion, an incredible number given that it was now worth more than Sears, Roebuck & Co.
With its coffers full and high-priced stock at its disposal, Akamai began to fill out its business through external means. In early 2000 it paid $200 million in cash and stock to acquire Network24 Communications Inc., a Cupertino, California-based streaming media company, the addition of which helped Akamai to move toward offering live Internet broadcasting services to additional web sites. A few weeks later, Akamai engineered a much larger deal, agreeing to trade $2.8 billion in stock to pick up InterVU Inc., another streaming media company. Until this time, Akamai was limited to working in Apple's QuickTime stream format, so that customers working in another format would have to provide Akamai with a streaming media file. The addition of InterVU allowed Akamai to create streaming media files in Real Networks' popular RealVideo and Microsoft's Windows Media formats, thereby broadening the company's customer base. Later in 2000, Akamai completed another strategic acquisition, paying $6 million in cash and stock for Santa Clara, California-based CallTheShots, Inc., a private company developing technology that would allow web sites to deliver customized content to users without incurring the expense of a larger infrastructure.
The price of Akamai stock peaked on December 31, 1999, at $344.88 a share. The outlook for the company continued to look rosy in 2000, but as the tech sector began to sour, along with the economy in general, Akamai got caught up in the downward spiral when the dotcom bubble began to burst. The company had quickly ramped up its hiring and bought a lot of real estate, only to see a large percentage of its customers going out of business, unable to afford its service, or delaying investments. During the first half of 2001 Akamai lost more than 12 percent of its recurring contracts. The price of its stock began to plummet as a result, and the company was forced to make deep cuts in staffing and to terminate leases. Akamai shares closed 2000 trading in the $21 range, a drastic change from the same time the prior year, but an enviable price compared with what was to follow. Akamai eventually bottomed out in 2002, worth little more than 50 cents per share.
In 2001 Akamai recorded sales of $163 million and a net loss of $2.4 billion, albeit the bulk of that loss was the result of restructuring and other charges. With those items excluded the company's loss totaled $429.1 million. The year also was marked by personal tragedy. On September 11, 2001, Lewin was a passenger on one of the airliners that terrorists crashed into New York's World Trade Center. Ironically, the act that cost his life also demonstrated the value of the Internet trafficking system he helped to develop, as the Internet was besieged with requests for information from people keeping track of events and Akamai's customers were able to cope with the spike in demand. Although the company continued to lay off staff in the remaining months of 2001, its performance on September 11 led to an increase in customers and played a significant role in Akamai beginning to rebound in 2002.
Akamai launched a new product in 2002, EdgeSuite, and began targeting large companies interested in establishing a significant web presence. EdgeSuite was essentially a more robust version of FreeFlow: combining the earlier dynamic traffic control features with added security features, increased speed on digital downloads, and the ability to deliver streaming media. A rash of new customers responded, including FedEx Corporation, Staples, MTV, and The Bombay Company. Akamai was also successful in attracting government customers, selling EdgeSuite to the Federal Bureau of Investigation (FBI), whose web site had experienced an increased level of traffic after September 11. As a preferred government vendor, Akamai also began selling its services to other federal government customers, such as the U.S. House of Representatives, who used Akamai to stream hearings on the Web. Despite this success, there was still more bitter medicine to be swallowed in 2002, as the company continued to shed workers. Employment that crested at 1,300 in 2000 stood at 550 at the end of 2002. The company was also the recipient of unwanted publicity when it was disclosed by the Boston Globe that Akamai was using college servers to deliver content from teen-pornography web sites and help offshore gambling sites to speed content delivery. Few people had realized that Akamai's system of spreading Internet content to geographically advantageous servers relied on college servers as well as company-owned servers. Schools benefited from the relationship because they gained faster Internet access and paid less money for more bandwidth. At first Akamai downplayed the controversy, maintaining that pornography and gambling represented less than 1 percent of its revenues and that it was not going to renew contracts with online gaming companies and no longer sought new business from what a spokesperson referred to as "adult content sites." Within a couple days the Boston Globe reported that the objectionable material was no longer mirrored on the college servers it checked.
Akamai's comeback continued in 2003, as reflected by the rising price of its stock, which traded below $2 at the beginning of the year and above $10 at the end. By February 2004, the price reached the $16 range, so that over a five-month period Akamai increased its market capitalization by $1.4 billion. The reasons for renewed investor confidence included Akamai's cost-cutting measures, increasing sales, and a restructuring of debt that took place in 2004. The outlook was also brighter because the market was beginning to come to the company, which had been criticized for overbuilding a network, intended for a much larger Internet of the future. But with the rising use of streaming technology to deliver music, movies, sports, and software, Akamai now found itself well positioned to service customers looking to take advantage of this demand. In 2004 Akamai became profitable for the first time, posting net income of $34.4 million on sales of $210 million.
One nettlesome concern for Akamai that continued to trouble the company in 2004 was its involvement in costly litigation to protect intellectual property. Initially Akamai fought with competitor Digital Island, and the matter continued with Cable & Wireless after the latter acquired Digital Island. Akamai won a patent infringement suit in 2002. But in that same year, Akamai became embroiled in a patent dispute with Santa Clara, California-based Speedera Networks, with both parties trading lawsuits and countersuits. Akamai accused Speedera of stealing trade secrets, while Speedera called Akamai a legal bully. In March 2005 both sides were weary of the distraction and the legal costs and, after six months of negotiations, agreed to settle their differences by joining forces. In a $130 million stock deal, Akamai acquired Speedera and quickly began to merge their operations and grow the combined company. In an odd way, the litigation was a benefit. According to Speedera's CEO Ajit Gupta, "We have been joined at the hip through litigation, so we know each other very well."
Principal Subsidiaries
AKAMAI Ltd.; AKAMAI GmbH; AKAMAI SARL; AKAMAI JAPAN; KAHUA HK Limited.
Principal Competitors
Kontiki, Inc.; SAVVIS, Inc.; Communications Corporation; Cable and Wireless PLC.
Further Reading
Greene, Tim, "Start-Up Seeks to Unclog Web," Network World, January 18, 1999, p. 6.
Lee, Dan, "Akamai Buys Rival Speedera," San Jose Mercury News, March 17, 2005.
McLaughlin, Tim, "Akamai Founders Net Cool Billions," Boston Herald, October 30, 1999, p. 20.
Penenberg, Adam L., "Speed Racer," Forbes, September 20, 1999, p. 200.
Poe, Robert, "Helping Web Sites Dish It Out," Upside, March 2000, p. 80.
Soule, Alexander, "With Akamai Rebound, CEO Is in Line for Stock Windfall," Boston Business Journal, February 20, 2004, p. 1.
Walker, Leslie, "Akamai Strives for a Safer, Speedier Net," Washington Post, September 30, 2004, p. E1.
— Ed Dinger
