More Back Pages stories

On market consolidation and the search for intelligent life

In the past several months we’ve seen the graphics market consolidate in two major markets: PC graphics and handheld graphics. This is not a new phenomenon; it happens in all industries including the CPU, display, and game console industry, and it isn’t over yet. Furthermore, when companies are diversified the consolidation in one industry is blurred somewhat.

The most recent consolidations in the PC graphics industry have been ATI’s acquisition of the graphics group of XGI, and Nvidia’s assimilation of the graphics team from 3Dlabs. XGI was a spinoff of the graphics group from SiS, and an acquisition of Trident’s graphics group. It remains to be seen if SiS will re-enter the graphics market, although there are rumors to that effect. If they did, it would be in the integrated sector.

Creative Labs, which acquired 3Dlabs in Q2’02, shut down the high-end workstation GPU makers in Q1’05, and Nvidia picked up almost all of the team and established a new operation in Huntsville in the process. 3Dlabs was one of the pioneers of the graphics market, starting from an acquisition of DuPont Pixel Systems back in 1994, and DuPont began that group as a result of an acquisition of Benchmark Systems in 1991.

So now the PC graphics industry is down to five suppliers: ATI, Intel, Matrox (the granddaddy of the industry), Nvidia, and S3/VIA.

Intelligent life

The recent acquisitions, however, have not been to gain market share per se but more important, to gain engineers—intelligent life.

Other acquisitions by ATI and Nvidia in China, Finland, India, Russia, and Taiwan have been done to accumulate bodies. In previous commentaries you’ve heard me point out that the demand for engineering talent is almost insatiable, and that it is the defining difference between the successful and not-so-successful companies. Whereas it was a big deal for a graphics semiconductor company to have two design teams, they are now approaching five as product development time lengthens and product lifetimes shrink.

The question this kind of logic always begs is where does it stop? When can you have enough engineers, or (god forbid) too many? Believe it or not, there is an answer to that question: when you start losing market share and/or the overall market starts to decline. And the sad irony of that is that it is a deadly cycle: you don’t have enough engineers to get the next-generation product out in a timely manner, or you didn’t have enough of the right engineers to make the next product competitive, and so you start to lose market share and now those once precious engineers become a burden. At this stage the risk of a declining overall market seems pretty remote. Demand is high and even increasing, and with the double whammy that Moore’s Law brings of more power and lower prices, it doesn’t seem like we’ll hit an asymptote for some years. Which means the demand for semiconductor design and manufacturing engineers will stay high, which means we will see continued acquisitions and takeovers.

Unintelligent life?

Here’s some more irony. Engineers who couldn’t get a product out, a product important to the market (i.e., the right specs, price, performance), might be considered uninteresting to other companies. In fact, though, they are very interesting. The fault of company or product failure is generally laid on management’s desk.

There’s a lesser demand for managers these days, whereas in the go-go years of the Internet bubble, it was all about new presidents, vice presidents, and marketing people; seems we’ve got enough of them right now.

Company growth

ATI and Nvidia are each approaching $3 billion in annual revenues. They diversified somewhat, but it only amounts to 15% or less of their overall sales. Unlike some of the bigger IT companies (Oracle comes to mind), the graphics companies don’t make acquisitions to increase the sales size of their com-pany. Rather, they make strategic technologies acquisitions, and that just fits with their cultures—they are technology companies, not process or production companies. These companies get growth by growing the market, and one of the side benefits is they almost never have culture clashes with the groups they pick up because they generally share a common dream—more, faster, and better graphics.

Market growth

The net result is fewer companies are shipping more chips per company, due to the combination of an expanding, yet consolidating market.

return to top