An updated version of that chip, the EV79 will follow about 12 months later, HP executives said at a meeting this week with financial analysts. At that point, HP plans to shift Alpha into what it calls "maintenance mode," a move that will save the company hundreds of millions of dollars.
The updated Alpha schedule is a slight departure from HP's previous timetable. In May, HP hadthat the EV7 would debut in late 2002.
HP is retiring Alpha, along with its PA-RISC processor, in favor of building systems using Intel's Itanium chip, which is based on an architecture HP helped develop. By shifting to Itanium, HP stands to save significant development costs. However, HP is betting big that Itanium sales will replace the business HP will lose from Alpha and PA-RISC based systems.
Alpha and PA-RISC processors account for $5 billion in annual revenue, more than a quarter of HP's total enterprise systems business, according to Peter Blackmore, HP executive vice president.
The Alpha business will lose about $200 million this fiscal year, Blackmore said. Once the final chips are out the door, HP will be able to scale back its investment and shift to what he dubbed "maintenance mode."
"That will flow through to the bottom line," Blackmore said at HP'sthis week. HP said that the enterprise systems business will be less profitable than the company hoped this year, in part because of the costs of developing for three high-end chip families.
The shift away from Alpha has been planned for some time. In June 2001, just three months before the merger between HP and Compaq Computer was announced, Compaq plans to cease work on Alpha, transferring much of the technology to Intel.
Executives have said that HP would stop coming out with new Alpha chips around the 2004 time frame and then try to get customers to migrate to Itanium systems by 2004. Although HP plans to stop developing new Alpha chips, the company has said it will keep selling Alpha-based systems as long as there is customer demand, likely until 2006.
Digital Equipment developed the Alpha in 1992. While touted for its performance, the chip and servers containing it never achieved the same sort of widespread acceptance in the marketplace as competing products from Sun Microsystems, HP or IBM. Microsoft developed a version of Windows for the chip, but then canceled it because of slow sales.
Compaq hoped to reverse the trend when it bought Digital in 1998, but gradually divested itself of the assets and employees behind Alpha.
Although a commercial flop, Alpha's influence remains wide in the industry. Many Alpha engineers have ended up at Intel and Advanced Micro Devices, and a number of the concepts earlier championed by Digital researchers, such as multi-threading, are embodied in the latest Pentium 4 and AMD's upcoming Opteron processor.
Large research institutions have been some of the biggest supporters of Alpha. Los Alamos National Laboratory uses Alpha chips in itssupercomputer, a mammoth system with thousands of processors designed eventually to perform 30 trillion calculations per second to simulate nuclear weapons explosions.
News.com's Michael Kanellos and Stephen Shankland contributed to this report.