The Development of the Xbox

Microsoftís Xbox project began in early 1999 when a group of Microsoft employees with videogame backgrounds began to explore the possibility of creating a game console. At thе time, Microsoft offered PC-based games, but personal computers had to be able to handle a wide range of applications, and thus were not optimized for videogaming.

While the gamers were trying to get Microsoft to develop a console, Microsoft management was also concerned that consoles such as the PS2 threatened the company ís home PC business. Developing a high-performance console, with a considerably lower sales price than a PC, might solidify Microsoft ís position in the home.

The group of gamers (referred to as ìrenegadesî by Dean Takahashi in his history of the Xbox project, Opening the Xbox: Inside Microsoftís Plan to Unleash an Entertainment Revolution) worked on their own to define a game console, including meeting with potential critical component suppliers. They made their initial proposal to Bill Gates on May 5, 1999, including a preliminary estimate of parts costs. The proposal was for a product that would leapfrog Sony, offering superior performance, and including a built-in hard disk drive that would enable online gaming and downloading of games and game enhancementsóin contrast to the PS2, which did not offer a hard drive, but at the time was considering one as an add-on. The console would also include a broadband Internet connection, despite the fact that broadband was available at relatively few homes at that time. The early plan anticipated sales of 1.8 million units in 2000, growing to 30 million in 2005 (at which time Microsoft would have a 35 percent market share)4 The project was legitimized as a business by the appointment of an executive to lead it in July 1999.

The Xbox team included potential suppliers in its planning from the beginning. Flextronics, a potential contract manufacturer of the Xbox, provided cost estimates and advice long before contract negotiations began. Nvidia, a leading developer of graphics chips, also provided cost and technical advice. In considering potential manufacturers, the Xbox team approached PC companies, such as Dell, since the Xbox was basically a computer. Discussions with the computer makers made it clear that the computer business model was inconsistent with manufacturing game consoles. Console makers typically lost money on each box, but made profits on selling games. Cutting prices on consoles made them more attractive to consumers, driving more game sales. Computer makers, on the other hand, had to make money on hardware sales. As Michael Dell explained:

When Sony cuts prices on its PlayStation, their stock price goes up. Every time I cut prices, my stock price goes down. If you donít understand why that happens, you donít understand the console business. I understand why [the Xbox] is strategic to Microsoft. I donít understand why [it] is strategic to Dell.5

Two factors were critical to profitability in the console business: driving costs down, and game sales. Microsoft estimated that the original PlayStation cost Sony $450 when it was launched, but just $80 five years later. Sonyís initial sales price had been about $300, decreasing to $99 after five years.6 Unlike computers, where performance typically increased dramatically, but prices stayed relatively constant, consoles maintained constant performance while costs decreased. As the project line matured, the loss on each console decreased, and eventually the manufacturer generated a profit on console sales.

This model fit well with the use of contract manufacturers to produce the consoles, and Microsoft began discussions with potential suppliers, eventually selecting Flextronics to build the consoles.

The Xbox would have over 1,000 components, of which about 45 were critical, either because of cost or because they were only available from single suppliers. One of the most important choices was the graphics processor, which was essential to the product performance. Sony had spent $1 billion to build a plant to manufacture its own graphics chip, enabling it to control its own destiny and reap all the benefits of future cost reduction.7

Microsoft leaned toward purchasing graphics chips from Nvidia, the preferred choice of game developers, despite a relatively high cost. In addition to considering other graphics chip companies, Microsoft also considered having its WebTV unit license technology and design the chip in-house. The chip could then be manufactured by a contract chip fabrication company, allowing Microsoft greater control over costs. Microsoft went so far as to promise a substantial investment in a new graphics chip company, GigaPixel before finally deciding to buy chips from Nvidia. (Microsoft did follow through with the investment, however.) They chose Nvidia because they believed that a reliable supply was more critical than obtaining the lowest possible cost. Nvidia had the best performance, and an excellent record of on-time development and delivery. GigaPixel was unproven, and did not have the available capacity to develop the chip. The eventual Nvidia contract included a maximum margin per chip, and decreasing maximum prices based on cumulative volume.8

Another critical component was the microprocessor, for which Intel and AMD were the competitors. AMD had been the front-runner through much of 1999, as it was eager to increase capacity, while Intel had dominated the market and took a hard line on prices. However, during the year, AMD introduced the 600 MHz Athlon processor, which did very well in the market, using up all of AMDís capacity. In order to make chips for the Xbox, AMD would need to add capacity, and wanted Microsoft to provide about $400 million for that purpose, plus a $300 million guarantee on chip purchases. AMD also increased its price. On the other hand, Intel had available capacity, and came down in price (though not as low as AMD), with prices declining each year. In March 2000, Microsoft signed with Intel to provide chips for the Xbox.9

In addition to manufacturing the Xbox, Flextronics also provided mechanical designers that helped with parts specifications. Parts procurement was handled in one of three ways. Commodity parts were purchased directly by Flextronics, which often used the same parts for many OEM products, and thus purchased in very high volume. For some custom designed parts, such as the controller, Microsoft negotiated the contract with the vendor, and Flextronics ordered the parts against the Microsoft contract. For parts such as the microprocessor and graphics chip, Microsoft purchased the parts from the vendor and delivered the parts to Flextronics.10

In December 1999, Microsoft decided that it would delay the launch of the Xbox from the fall of 2000 to the fall of 2001. A 2000 launch would force the company to use technology that did not offer a substantial improvement on Sonyís PS2, which was scheduled for launch in spring 2000. By waiting, Microsoft would be able to incorporate much faster graphics and microprocessor chips, offering dramatically better performance than the PS2. The delay, however, added the risk that Sony would solidify its position, making it harder for the Xbox to gain market acceptance.11

When the Xbox was launched in November 2001, it used Intel ís 733 MHz Pentium III chip as its central processor, with a 250 MHz Nvidia graphics processor. The Xbox also had a DVD player, 64MB memory, an 8 GB hard drive, and built-in broadband modem for Internet access.12 The system processed 116 million polygons per second, the critical measure of graphics performance, compared to 66 million for the PS2.13