IBM Unveils the Personal Computer at New York's Waldorf Hotel
IBM launched the 5150 in August 1981 with an unusually modular design whose documentation and suppliers helped create the PC-compatible market.
On August 12, 1981, IBM introduced its Personal Computer, model 5150, at New York’s Waldorf Hotel. The base machine was advertised at $1,565 with 16 kilobytes of RAM and no disk drive or display. That price made for a memorable headline, but it did not describe the configuration most business users would actually want. Adding a monitor, diskette drives, printer, more memory, an operating system, and applications raised the total substantially.
That difference is important when reconstructing the launch. IBM promoted software such as VisiCalc and EasyWriter alongside the PC, but the $1,565 entry price should not be interpreted as a complete office system with those applications bundled at no additional cost. The low figure represented a usable starting point for expansion, not the typical all-in purchase.
A one-year project inside a much larger company
IBM did not invent personal computing in 1981. Apple, Commodore, Tandy, and a community built around S-100 systems had already established a market. IBM’s challenge was to enter quickly without subjecting the product to every stage of its traditional mainframe development process.
William C. Lowe, who ran IBM’s Boca Raton laboratory, proposed building the machine from available components through a small, semi-autonomous effort called Project Chess. Lowe initially led the work; after he moved to another IBM position, Philip “Don” Estridge took charge of the team that completed and launched the PC. Treating both men as simultaneous directors throughout the project blurs that handoff.
The team chose an Intel 8088 processor, clocked at 4.77 MHz, and relied heavily on components from outside suppliers. Five internal expansion slots allowed buyers and third parties to add memory, display adapters, storage controllers, communications hardware, and other capabilities. The system included BASIC in read-only memory, so even a diskless configuration could do something useful when connected to a display.
“Open architecture” needs a qualification
The 5150 is often described as an open-architecture computer. In a practical 1981 sense, that description fits: IBM published extensive technical documentation, used recognizable off-the-shelf parts, exposed expansion interfaces, and encouraged an ecosystem of hardware and software around the machine. Those decisions made the design easier to understand and extend than a sealed, entirely proprietary system.
It was not “open” in the modern open-source sense. IBM’s BIOS firmware was copyrighted, and a legal compatible maker could not simply copy it. Clone manufacturers eventually used clean-room engineering to reproduce its behavior without reproducing IBM’s code. That technical and legal boundary is why compatible BIOS implementations became so important to the clone market.
The operating-system story also deserves precision. IBM offered Microsoft’s PC DOS as the least expensive and ultimately dominant choice, while UCSD p-System and CP/M-86 were also options. Microsoft retained the ability to supply its related MS-DOS system to other computer manufacturers. The combination of broadly available hardware, documented interfaces, and an operating system that could appear on non-IBM machines mattered more than any single decision in isolation.
What buyers saw in August 1981
IBM’s name reassured business customers who might have regarded earlier microcomputers as hobbyist tools. Its dealer network, service capacity, documentation, and applications made the product legible to corporate purchasing departments. The “Little Tramp” advertising campaign, inspired by Charlie Chaplin’s screen character, presented the PC as approachable rather than as a miniature mainframe reserved for specialists.
The machine’s specifications were modest even by later standards, but they were expandable. Configurations could grow from 16KB to 256KB of memory on the system board and expansion cards, use cassette storage or 5.25-inch diskette drives, display text or color graphics, and attach a printer. VisiCalc supplied a compelling spreadsheet use case; EasyWriter provided word processing. These applications were part of the launch ecosystem, not proof that every base box included a ready-to-run office suite.
IBM’s own historical account says more than 750,000 PCs were sold within two years, far beyond internal expectations. That documented two-year figure is more defensible than repeating loosely sourced claims about exact sales in the first four months or orders by Christmas. The commercial result demonstrated that a major enterprise-computing vendor could legitimize the personal computer for offices at scale.
The standard escaped IBM’s control
The lasting impact was not merely that IBM sold many 5150s. Other companies built expansion cards, peripherals, applications, and eventually machines compatible with the same software environment. Once legally engineered compatible BIOSes became available, Compaq and other vendors could compete while preserving the value of customers’ DOS software and accessories.
That network effect shifted control of the platform. Intel supplied the processor family, Microsoft supplied the dominant operating system, and clone makers competed on price and performance. IBM later tried to regain architectural control with the proprietary Micro Channel bus in its PS/2 line, but the wider compatible ecosystem resisted the change.
The careful conclusion is not that IBM altruistically gave away a fully open design, nor that the PC became a standard by accident alone. A compressed schedule encouraged modular sourcing and unusually detailed disclosure; IBM’s market authority then drew buyers and vendors to those choices. Together they produced the IBM-compatible platform whose descendants shaped personal computing for decades.
Related:
- The IBM PC: How One Product Accidentally Created an Industry Standard
- From Mainframes to Microprocessors: The Decades-Long Shrinking of Computing
Sources: