High Hopes
Silicon Graphics aims to repeat its success in the graphics marketwith a low-end workstation. But, says Sean Hallahan, it must firstovercome stiff competition and some dodgy chips
Silicon Graphics has been forced to issue a profit warning for its first quarter because of the need to replace thousands of faulty chips.
Problems were found with the R10000 chipset which lies at the heart of the company's products and Silicon Graphics (SGI) was forced to recall thousands of machines between March and July this year.
The news of the profit warning and the chip failure which caused it has come as a double blow to the company, which has just released its range of low-end workstations - the O2.
SGI moved quickly to allay the fears of its resellers and customers by suggesting that the problems that occurred with the R10000 chip were not uncommon among other semiconductor manufacturers. The company also claimed that the new machines would not be affected by the problems with the R10000. 'It was a one-off manufacturing problem which we have now resolved and the machines have been, or are being, replaced,' says an SGI representative.
The representative went on to point out that the profit warning was partially a result of SGI's phenomenal rate of growth in previous years which could not be sustained indefinitely. 'We did see a stall in sales while customers were waiting for new products to arrive. We have not had the same rate of growth as in the past but we are still growing ahead of the market,' hesays.
But SGI's troubles will almost certainly be greeted with some delight by its rivals, most particularly Sun Micro-systems. Although both companies operate in the workstation arena and both share a common heritage in US Academia, they have evolved in very different ways.
Stanford University in the US was the cradle for two of the IT industry's major success stories - Sun Microsystems and Silicon Graphics. Both companies were started in 1982 and their founding fathers were researchers at the university. Sun made its presence felt, initially in the scientific and engineering workstation sector, but it later achieved a considerable measure of success in the financial market. Silicon Graphics concentrated much of its efforts on developing the graphics workstation and took a more niche approach to marketing than its sibling and rival.
SGI recognised that it was weak in the markets where Sun was strong - the financial and commercial sector. In December 1995, it appointed its first UK distributor, Access Graphics, with the aim of broadening its approach to the commercial market. Access hoped to recruit 130 dealers during the course of 1996, but the problems with the R10000 chip will sound a note of caution in the channel. According to the SGI representative, the O2 is a machine firmly aimed at the volume market. Currently the split between direct and indirect sales is 50/50 but SGI aims to bring that to a 60/40 split within a matter of months.
Silicon Graphics built its systems around the Risc (Reduced instruction set computer) chips from Mips. In 1992 when Mips, after a massive and costly attempt to break into the commercial market, hit financial problems, it merged with SGI. The company has clearly learned from the mistakes made by Mips and has concentrated its efforts on the area it knows best, powerful graphics workstations. In 1995 it merged with two other companies, each a specialist in its field, Alias Research and Wavefront Technologies.
A new subsidiary was formed to market the powerful graphics processors to the entertainment industry.
Wavefront concentrated on the film business and provided much of the graphics technology behind Star Trek and Jurassic Park while Alias was chosen for the development of the Nintendo Ultra 64 games machine. Silicon Graphics also pulled off another major coup when it acquired Cray Research, the supercomputer supplier.
But it is not just Sun that Silicon Graphics will face as a competitor in the low-end workstation market. Although Sun has the largest market share, according to analysts, Hewlett Packard is also active in the field.
In addition, SGI faces a challenge from another newcomer to the workstation market and one with a considerable pedigree in both the technology and in terms of channel marketing - Compaq. Earlier this year Compaq set up its own workstation division and is promising to launch workstation products before the end of the year. It is difficult for any newcomer to break the stranglehold of the established market leader, particularly when, like SGI, it is coming late into the field.
Compaq, the PC market leader, can play to several strengths, its number one position as supplier of PCs, its acknowledged technological excellence and the fact that it has had a reseller channel established for well over a decade. This latter point may prove decisive when the battle of the workstation suppliers takes off. Other companies like Hewlett Packard and Digital can ride on the back of their server offerings and installed base to market their own brand of low-end workstations.
SGI has no such advantages. Its reseller chain is still being established and, since it has virtually no presence in the commercial arena, it has no ready made installed base in which to market low-end machines.
Add to these facts the double whammy of the R10000 problem and the posting of profit warnings and the future of SGI in the low-end market begins to look very weak. Even if the R10000 problems have no bearing on the machines to be announced there will be, nonetheless, those who will treat the arrival of the boxes with some caution.
'Compaq is undoubtedly a force to be reckoned with but it does not have our ability to innovate,' the SGI representative said. He explained that the SGI strategy was to innovate at the high end of the product range and then filter the technology down to the low-end systems.
Sun is already taking advantage of SGI's dilemma. 'We have noted their trouble and we are targeting both their resellers and their installed base because we think they are weak at the moment,' says Chris Sarfas, Sun desktop marketing manager.
Sarfas identifies what he believes is a further problem for SGI - the integration of Cray with the rest of the business. 'As we know from the past, such things are never easy,' Sarfas says.
He believes that SGI's move into the lower end of the market is governed by the fact that its existing markets are saturated. 'SGI, in the niches where they are strong, are very strong but the problem is those niches are not big enough,' he says.
Sarfas points out that there are very few instances of companies in the IT industry coming late into a market and acquiring any appreciable market share. Digital is a prime example. The company entirely missed the significance of the Intel Dos-based platforms and produced a range of personal systems that were not IBM compatible. Not until several years later did Digital deign to bring out a compatible machine, by which time it was too late.
Mike Clarke, business unit manager for Access Graphics, does not believe that the problems with the R10000 will adversely effect the sales of the low-end workstation, the O2. 'The business of the R10000 was handled quite well. Silicon Graphics acknowledged the problem and informed everybody straightaway,' he says.
Access is well on its way to meeting the target of 130 re-sellers, according to Clarke, with 85 already signed up and the remainder in a period of transition. Access Graphics has already taken delivery of over 100 of the new systems and Clarke says that there is a great deal of interest from the channel. The O2 workstation currently uses the R5000 processor which was not faulty but the company plans to introduce R10000 boxes in the future. SGI claims that all the problems with the R10000 chip have been sorted out but admits that this has been at considerable expense to the company.
Access Graphics is certain that the O2 will be a winner. 'It is a box for all markets, ideally suited to the studio, multimedia and Cad. This is the first true volume product that Silicon Graphics has ever produced,' says Clarke.
'People are asking for demonstrations and are asking to buy the machines and that is unusual. SGI resellers would normally ship a product against an order. We recently did a like-for-like comparison for a Cad dealer running the O2 R5000 against a PC running NT and it outperformed it. The dealer was asking why he should bother with a PC and NT for CAD applications.'
There is one aspect of SGI's channel strategy that may attract dealers.
The company is determined to keep the reseller margins high - far better than in the cut throat PC sector, according to Clarke.
For the specialised markets like Cad and entertainment this may well be true, but if SGI wishes to break into the commercial sector as Sun has done it is going to have to dirty its hands and involve itself in a price war. Currently, the O2 R5000 costs around u5,000. But with companies like Compaq slashing PC prices by up to nearly 60 per cent on some models, many dealers may prefer to make their margins from volume sales rather than the high profit but slower moving workstations.
It is still too early to know whether SGI will be able to extend the market for the low-cost workstations beyond its traditional niche markets.
Certainly one of those niche markets is Cad from where much of the early interest in the O2 has come and it is not surprising that companies operating in the field are prepared to look for a powerful alternative to the likes of HP, Sun, and Digital's Alpha.
But the commercial market is a jungle compared to the markets in which SGI has been operating. SGI has been extremely successful in its niche areas, but then it has had to face hardly any serious competition. While Sun and other companies have made token forays into the territory dominated by SGI they have not been overly successful. Now, companies like Sun will not yield their markets to SGI lightly.
Compaq's entry into the workstation arena will also make life complicated for SGI and its resellers. There is no denying that, problems with the R10000 notwithstanding, SGI's technical reputation is very good. But then so is Compaq's and Sun's.
While it is true that Compaq does not come from a traditional workstation background and that it will need to acquire expertise in the area it does have some advantages over SGI. Of these, its channel strategy and its installed base in the corporate PC market are critically important.
Initially at least, SGI will be pushing its machines into areas where it is already strong. 'We are targeting the creative professional in product design. In the entertainment business, where we are already known, the thing that has stopped us selling more machines was simply that the customer could not afford to hire the people to sit in front of the screen because the capital cost was so high,' the SGI representative says. But the O2 provides functionality at u5,000 that would have cost u20,000 to u30,000 12 months ago, he adds.
Whether SGI can break out of its niche markets with the O2 is a moot point. But in the first instance it seems that its plan is for the low-cost, low-end machines to widen its existing niche markets. If it confines itself to those niches it should prove successful, but moving into the wider commercial world will almost certainly instigate a battle royal.