AS/400 Business Partners and customers alike are increasingly concerned about the rapid, dramatic, and somewhat unexpected decline in AS/400 server sales in 1999. IBM rarely gives specific revenue figures or revenue growth rates when it announces its quarterly financial results, but in the third quarter, IBM’s Chief Financial Officer, Douglas Maine, came right out and said that AS/400 sales for the quarter were down a stunning 30 percent from last year. While AS/400 server sales had been down all year—about 11 percent in the first quarter and about 12 percent in the second quarter—very few people, including executives in the AS/400 Division and its parent Server Group as well as Wall Street analysts, expected such a big drop in AS/400 sales. The problem is that things will apparently get a lot worse before they get a little better, and this has serious implications for Business Partners who depend on the AS/400 for their living and AS/400 customers who depend on it for their data processing.
Two factors seem to have caught IBM off guard this year. First, IBM had a stunning year in 1998 after a killer year in 1997, and it has been very difficult for Big Blue to separate AS/400 customers’ enthusiasm for native enterprise resource planning (ERP), client/server, e-business, and groupware applications from processing capacity demands driven by Y2K remuneration and testing. IBM has been surveying its server customer bases for years, trying to gauge server buying for Y2K projects so it could assess what effect, if any, the end of the century would have on demand for AS/400s and other servers. In July, when IBM announced its second quarter results for 1999, CFO Maine confidently told Wall Street analysts that IBM was not yet seeing any slowdown related to Y2K and said that the company did not anticipate any slowing for the remainder of the year, based on its most recent surveys. These surveys, it turns out, were apparently wrong. For the first two quarters of 1999, IBM has blamed the AS/400 server sales downturn mostly on problems in Europe (the largest AS/400 customer base) and “sales execution” and “model mix” problems throughout the world. While these explanations are certainly valid—model 170 Invader servers are powerful enough for many AS/400 shops, and customers buying these boxes, which have serious capacity limitations compared to regular AS/400e 7xx servers, tend to spend less money—it now looks like IBM has grossly underestimated the amount of processing capacity AS/400 customers bought in 1997, 1998, and early 1999 to prop up Y2K projects, which will be used in the coming quarters to run production work now that Y2K is essentially over. The upshot is that it is difficult to imagine the AS/400
market rebounding much before March 2000, and perhaps later, depending on how much extra processing capacity customers have laying around that they can use to handle extra demands on their legacy applications and run other, modern workloads.
IBM, of course, doesn’t provide any specific quarterly numbers to give AS/400 customers and Business Partners a handle on what’s going on in its server markets. John Jones, the head computer analyst at Salomon Smith Barney, has the best and most detailed estimates of IBM’s server sales among the leading Wall Street brokerages. According to his estimates, which are detailed in Figure 1, IBM will sell about $500 million in AS/400 servers in the fourth quarter, down 47 percent from last year’s $935 million. Jones figures that IBM will be lucky if it sells $2.5 billion in AS/400 servers in 1999, which would be down 26 percent from the $3.3 billion that he reckons IBM sold in 1998. That would make 1999 the worst full sales year in the history of the AS/400 line; in 1988, when IBM had only five months of AS/400 sales, it pulled in nearly $2 billion in processor sales. The whispered message from IBM’s CFO, by the way, is that the sales lost in 1999 are gone forever, that there will not necessarily be a corresponding pickup in demand in 2000 equal to the $835 million in AS/400 server sales IBM didn’t make to match 1998’s results.
No one is willing to venture a guess about 2000 sales yet because there is another AS/400 product transition coming when the I-Stars start shipping, probably around May. And with AS/400 sales so dependent on sales of ERP software suites, it is even harder to make any reliable guesses. If 1999 saw a slowdown as Y2K approached, 2000 will see a speedup as it recedes, and the two years will very likely look like mirror images—meaning sales of about $2.5 billion in 2000, with most of those sales in the third and fourth quarter—unless IBM does something drastic to stimulate demand. With sales down so far, I doubt IBM will be tempted to drop AS/400 prices to stimulate demand, but I think that if it prices the I-Stars aggressively enough, it could help it expand into new markets faster than it might otherwise and, in the long run, give it a bigger installed base from which it will get a larger and more predictable demand for increased processing capacity. The trick, as IBM well knows, is getting customers who are sick of NT and OS/390 and baffled by UNIX to buy packaged solutions on an AS/400. As it now stands, the AS/400 is too expensive a platform, regardless of reliability and ease of use, and too hard a sell to be a high-volume product. While IBM is not saying that this is a cause of its recent AS/400 sales downturn, I am, and the customers I am speaking to are as well.
This is an important idea, and it is one with which many AS/400 Business Partners disagree, vehemently at times. If IBM cut AS/400 prices to make the platform more attractive, AS/400 resellers would have to sell all that much more equipment just to keep running in place. With about three-quarters of AS/400 sales going through the BP channel (by revenue), a big price change would have a dramatic ripple effect on their ability to meet their payrolls. With AS/400 revenues off so sharply, it isn’t hard to see why many BPs were arguing for IBM to increase AS/400 upgrade prices. (See the related story “IBM Raises Upgrade Prices and Blood Pressure Levels” on page 49 for more on that). But this price increase goes against the general trend in the midrange arena and is detrimental to BPs that are trying to attract new customers to the platform. That is where IBM’s focus should be. With Compaq selling 1 million PC servers a year, no one can tell me that there is not an opportunity to sell hundreds of thousands of AS/400s a year at a modest premium to those Compaq servers, because the AS/400 is a better box. That is what will make the AS/400 a $5 billion-a-year business, and yes, it will hurt Netfinity sales. Too bad.
The AS/400 is not the only IBM server line that has Y2K lockdown sales problems. IBM’s S/390 mainframes have also been hit very hard. IBM said that S/390 MIPS shipments were down 18 percent after five consecutive quarters of MIPS growth (averaging more than 100 percent) and that revenues from these MIPS sales were off 40 percent. That puts S/390 server sales in Q3 at about $805 million, significantly off from the $1.3 billion IBM got after announcing the G5 mainframes in last year’s third quarter and considerably lower than the $925 million the company pulled in from this year’s second quarter in the wake of the G6 mainframe announcements. Still, Wall Street was expecting
mainframe sales more in the neighborhood of $850 million to $900 million. For the year, SSB’s Jones expects that mainframe sales will be down 24 percent to $3.25 billion, the worst year in IBM mainframe history.
The RS/6000 and Netfinity server lines have, apparently, been less insulated from the chilling effects of the Y2K slowdown. The overall RS/6000 server business looks as if it will grow by about 4 percent in 1999 to $2.6 billion, according to Jones’ estimates. According to my estimates, Northstar-based RS/6000 SMP server sales account for about $2.2 billion, or about 87 percent, of total RS/6000 sales for the year (up from 64 percent of total sales in 1998) and are growing at close to 30 percent in the last two quarters of 1999. But this growth is not enough to offset declines in sales of RS/6000 SP parallel servers or Power3-based technical servers. Netfinity server sales look as if they will grow to about $2.35 billion for the year, up 63 percent, just shy of AS/400 and RS/6000 server sales. The real stunner is that it appears as if Netfinity server sales will eclipse AS/400, RS/6000, and S/390 server sales during the first quarter—which would be the first time that has ever happened and, unless IBM changes its tactics to protect its software franchise, probably not the last time. The upshot, as Figure 1 shows, is that no matter what happens with the Netfinity line, the combined AS/400 and RS/6000 Apache and Northstar server businesses generate well over $1 billion per quarter in revenues for IBM, which is as much or more than its S/390 server business. Northstar is very, very important to IBM.
Many believe that a rebound in the AS/400 server market will be driven by a rebound in ERP software sales once companies get beyond the millennial change. As Figure 2 shows, according to market researchers AMR Research, the ERP market is expected to rebound after a difficult 1999. But a rebound in AS/400 sales may not necessarily pan out. In 1996, a combination of a number of different effects—mainly lower hardware and software costs to make buying ERP suites attractive—drove the ERP software market up to $7.2 billion, according to AMR Research. In that year, AS/400 server sales were approximately $4.5 billion, due mostly to pent-up demand for RISC- based AS/400s. In 1997, the ERP market grew by 66 percent to $11.9 billion, driven primarily by many companies deciding to get out of their Y2K tangles by buying ERP suites rather than fixing legacy code. But AS/400 server sales fell 27 percent to $3.3 billion. As ERP sales grew by 39 percent in 1998, AS/400 server sales were once again flat at $3.3 billion, even after a $500 million boost from Domino-related sales. This implies that AS/400 sales related to ERP fell, even as ERP software sales were going up. And this year, with ERP sales growing by a more modest 22 percent to $20 billion, AS/400 sales will be off about 26 percent to $2.5 billion. My best guess is that AS/400 server sales will be up 5 percent in 2000, even as ERP sales resume their 37 percent growth rate. IBM will have to do something drastic to get AS/400 server sales back up to their regular $3.3 billion level and perform some miracles to get the AS/400 business growing at the same rate as the ERP market.
All is not lost, however. If IBM and its software partners get supply chain management (SCM) and customer relationship management (CRM) software suites running natively on the AS/400, growth in AS/400 server sales could resume. These two markets are growing at 45 percent to 50 percent per year, but very few SCM and CRM applications are, as yet, available for the AS/400. Maybe IBM could stop spending billions on stock repurchases each quarter and pay i2 and Manugistics in the SCM space and Corepoint, Siebel, Vantive, and Clarify in the CRM space to port their code to the AS/400 and actually promote it.
1500
1,500 M1200
1,200
M
900 M
600 M
300 M
900
IBM Quarterly Server Revenue
1500
600
1200
300
900
Source: Salomon Smith
0
1Q98 2Q98 3Q98 4Q98 1Q99 2Q99 3Q99 4Q99
600
Apache/Northstar/Pulsar AS/400 Server Sales RS/6000 Server Sales Netfinity Server Sales S/390 Server Sales
$1,045 M $1,256 M $1,157 M $1,395 M $1,202 M $1,362 M $1,050 M $1,085 M $705 M $890 M $760 M $935 M $625 M $780 M $550 M $500 M $680 M $665 M $510 M $605 M $695 M $685 M $540 M $650 M $230 M $253 M $310 M $650 M $465 M $500 M $585 M $800 M $805 M $765 M $1,305 M $1,420 M $860 M $925 M $805 M $655 M
300
Figure 1: AS/400 and S/390 sales have taken a beating in 1999
0
70 B 60 B 50 B 40 B 30 B 20 B 10 B
ERP Sales SCM Sales CRM Sales
Source: AMR Research
1996 1997 1998 1999 2000 2001 2002 2003
$7.2 B $11.9 B $16.6 B $20.2 B $27.7 B $37.4 B $50.1 B $66.6 B $1.1 B $1.8 B $2.6 B $3.9 B $5.8 B $8.8 B $12.9 B $18.6 B $0.8 B $1.2 B $2.3 B $3.7 B $5.4 B $7.9 B $11.5 B $16.8 B
Figure 2: ERP software sales are expected to rebound in 2000, and SCM and CRM softwares sales are growing at 40 percent or more per year.
LATEST COMMENTS
MC Press Online