The Socket Wars How Intel is giving R&D a bad name

The Socket Wars How Intel is giving R&D a bad name

The Socket Wars How Intel is giving R&D a bad name


Hal Plotkin, Special to SF Gate
Monday, December 7, 1998

URL: http://www.sfgate.com/cgi-bin/article.cgi?f=/g/a/1998/12/07/socket.DTL

Years ago, most “research and development” spending went into the lab, giving us everything from light bulbs to integrated circuits to pacemakers. By focusing on the “Really Big Questions,” R&D often led to incredible, not just incremental, leaps.

Wall Street analysts began evaluating companies by measuring the proportion of corporate revenues devoted to R&D. And politicians enacted corporate tax credits to subsidize R&D. Lionized by the press, R&D became technology’s most sacred cow.

It’s time to take another look at where at least some of those cows are grazing. Today, sizable chunks of corporate R&D budgets are being hijacked by so-called strategic planning departments populated by nearly as many lawyers as engineers or scientists.

This is particularly true in industries dominated by a handful of major players. In those sectors strategic planning, and the research and development needed to bring those plans to life, often translates into figuring out ways to kill competitors and monopolize markets.

The latest example of this disturbing trend is something you might call the socket wars. Several companies, most notably Advanced Micro Devices, Inc. and Cyrix, are now making microprocessors (CPUs) that have essentially caught up with longtime industry-leader Intel.

In response, the Santa Clara-based behemoth’s research, development, and marketing team is attempting to pull the rug out from under them. Or more accurately, pull the rug out from under their motherboards.

Rather than just concentrate on building a better microprocessor, Intel is aiming to preserve its high profit margins by forcing a change in the CPU sockets used on PC motherboards. CPUs, after all, are useless slivers until plugged into a PC motherboard where, linked with other chips and controllers, they perform their magic. Intel’s much too clever strategic planners figure the best way to protect the company’s market share for CPUs is to control the designs for the sockets into which they are plugged.

It’s easier to understand all this if you think about how a similar strategy might work with more familiar consumer goods. Say, for example, you had a lamp maker with an eighty percent market share.

Worried about other companies horning in on the growing light bulb market, the monopoly lamp maker suddenly phases out the old lamps and patents new, better lamps with square sockets instead of round ones. As a result, no other companies are allowed to make light bulbs for the new lamps. And you have to throw your old lamps away because the newer, more powerful bulbs won’t work in them.

Intel’s marketing mavens hope most people won’t follow the technically arcane fight over motherboard sockets. Instead, they want consumers to think that if a Pentium is better than a 486, which it surely was, then a Pentium II-system, which happens to (shhhhhh!) contain the new proprietary socket, must be better than a plain old Pentium. It’s just the next step up the ladder. Right this way, folks! We’re here to help.

The problem with Intel’s approach is it threatens to deprive consumers of the benefits of competition in the microprocessor market right at the moment the fight has finally been joined. To be sure, Intel claims it moved to the new socket architecture primarily to improve overall system performance. It’s just an unfortunate coincidence that no competitors’ CPUs will work in Intel’s new sockets, called Slot-1.

Unlike Intel’s earlier, and largely successful, efforts to monopolize the CPU market, the company’s ploy might not work this time. Despite Intel’s claims, there are several very competitive non-Intel microprocessors already in the market.

Meanwhile, rival Advanced Micro Devices, Inc. has organized The Super 7 Platform Initiative, designed to make sure currently standard Socket 7-based products keep pace with, or exceed, Intel’s new Slot 1-dependent microprocessors.

The outcome of this skirmish is critical.

If Intel succeeds in making its proprietary Pentium II Slot-1 PC architectures the de-facto standard, efficiencies gained by Socket-7 system makers will be lost. Open standards-based Socket 7 motherboards, and chip sets designed to work with them, are now as cheap as pancakes, a fact that has helped to dramatically drive down the costs of PC ownership.

Personal computers, particularly those made by so-called clone makers, have finally become commodity products, with prices determined more and more by robust competition rather than the “what the market will bear” approach typical of Intel’s most monopolistic, early days.

For more than a decade, Intel has priced its microprocessors with little regard to what they cost to produce. Instead the question, taught in the finest of business schools, was: how much will the consumer pay? That’s why my old Intel 8088 PC cost roughly the same as my 286, 486, and Pentium systems. Intel had my number.

It’s a strategy, sometimes called market segmentation, Intel Chairman Andrew Grove recently explained, during a speech to the Academy of Management. Finding new ways to get top dollars from customers is always a popular topic at such gatherings.

But it is not a popular subject down in the aisles of most computer stores. Fortunately, some big players are not dancing to Intel’s tune. IBM, for example, while offering pricey Pentium II, Slot-1 systems, is also selling a lower cost, high performance Aptiva E4N, a Socket 7 PC with a souped-up 350 megahertz AMD K6-2 microprocessor.

Several other companies are likewise now selling inexpensive Socket-7 PCs, some for as little as $600. They don’t sport the little “Intel Inside” stickers. But if you buy them, they will leave you with several hundred dollars more inside your pocket.

It’s up to consumers to determine whether Socket 7 systems, and the competition they have created, survive now that Intel has abandoned them. If we’re lucky, Intel’s proprietary Slot 1 systems will whither into a dead limb on the PC’s evolutionary tree.

What bothers me most, though, is how unimaginative Intel’s strategic planning department has become. It doesn’t take much creativity to sabotage the other guy. With enough money, any bully can usually do it.

The innovators we remember most fondly, however, won their lasting reputations doing breakthrough science and pioneering revolutionary products, not by finding clever new ways to kill competitors and keep prices high.

About the Author /

hplotkin@plotkin.com

<p>My published work since 1985 has focused mostly on public policy, technology, science, education and business. I’ve written more than 600 articles for a variety of magazines, journals and newspapers on these often interrelated subjects. The topics I have covered include analysis of progressive approaches to higher education, entrepreneurial trends, e-learning strategies, business management, open source software, alternative energy research and development, voting technologies, streaming media platforms, online electioneering, biotech research, patent and tax law reform, federal nanotechnology policies and tech stocks.</p>