Intel: Who exactly is running the show?
Lights on, nobody home
Opinion Headless chickens are all very well if you're talking about one chicken that has unfortunately had its head chopped off and made into a pie. If you're talking about a multinational, multibillion dollar corporation employing more than 70,000 people, having no one in control is a tad more significant.
We were wondering today if there is anyone at a senior level within Intel who is actually exerting any real control over where the company is going. With every roadmap, press release and leak reaching us here at Vulture Central, it would appear that the answer is no. We think the major stockholders should be asking some serious questions of the men who claim to be running things.
Intel is now driven by roadmaps, PowerPoint presentations and hot air. Once upon a time it was run by engineers. Grove, Barrett, Gelsinger and Co still know how to design and build great semiconductors, so how come none of them is saying 'no' to the marketing department when they come up with their latest hairbrained scheme to launch a crappy chipset that doesn't work, or a processor that can't be delivered?
Intel was once a technology hot house, innovations and real quality products streaming onto the market. Other companies stood, slack-jawed, in admiration and played what seemed to be a never-ending game of catch-up. No wonder, then, that Intel referred to the competition as mere imitators.
But two years ago, cracks began to appear in the foundations of Intel's ivory tower. The debacle of the first Celeron (the cacheless Covington) was the first evidence that all was not well at Satan Clara. Why? Because Celeron was the first product that Intel had ever produced to combat a perceived threat from an imitator.
While Intel had blithely carried on launching mid range and server/workstation products, AMD had been winning market share from Chipzilla big time at the low end, especially in retail sales, and perhaps more significantly in the US.
The first time AMD issued a press release claiming dominance at the entry level, dusty and be-cobwebbed alarm bells started to ring at Intel.
This was unexplored territory for Chipzilla. Overnight it moved from a technology-driven innovator to a marketing-led decapitated fowl. Sure, it had lots of highly paid marketing and PR goons on the payroll, but up until now, they had had very little to do. Now they had to earn their money.
This they have singularly failed to do.
Intel has been drawn into a futile megahertz war with AMD. One can only applaud the marketing skill of Chimpzilla in pushing exactly the right buttons to make Intel look stupid. Intel busted a gut to demonstrate a 1GHz Pentium III to underwhelmed hacks at the last Developer Forum. AMD announced the availability of a GigaHertz Athlon just days before Intel launched the GHz Pentium III. You still can't get a 1GHz Pentium III, and the latest roadmap leak tells us you won't be able to until Q3 at the earliest.
This is marketing bullshit at its worst 'launch' a product in an attempt to steal the opposition's thunder in the full knowledge that you can't possibly deliver it for the best part of six months. Meanwhile, OEMs are up in arms because they can't get enough Intel parts no wonder they're turning to the AMD alternative.
(Challenge to Intel: prove us wrong deliver a working 1GHz part that isn't marked 'Engineering Sample' to us within 24 hours, along with a list of disties and OEMs who have received production parts)
Oh no, it's bloody Caminogate again
Look, we're really sorry. We know you're all bored to death with the trials and tribulations of the zombie bloodsucking chipset from Hell, but tough.
Rambus is a great idea, says the marketing department. But it doesn't bloody work, reply the engineers. How long will it take to fix, ask the marketeers. Eighteen months, reply the engineers, but we can do you an interim solution that'll use SDRAM. It'll run like a slug and we'd need at least a year to get it sorted. You've got six months, instruct the folks wearing white socks.
And blimey, the i820 gets launched, very late, and it doesn't work. It gets revamped to only support two DIMMS or RIMMS and it still doesn't work. The Register says the SDRAM versions are being recalled. Oh no they aren't, bleats Chipzilla, only to do a radical and expensive U turn four weeks later.
Intel once claimed it had learned a tough lesson when the original Pentium floating point bug the infamous FDIV flaw hit the fan. When a few scientists at universities around the world discovered that 2+2 equalled 3.0197886, Intel's reaction was to bury its head in the sand.
Only after public pressure increased to a point where the company actually had to do something about it were the faulty parts recalled. On the induction courses it runs for new employees, Chipzilla still puts up a slide simply reading 'FDIV' and explains that company policy is now to come clean immediately a problem is identified.
Explain then why the company tried to bluster its way out of the i820 farce by claiming there was nothing wrong. We discovered that a total recall of Cape Cod i820 mobos was about to happen and ran a story to that effect. Intel denied there were any recall plans and continued to deny it for a month until finally putting its corporate hands up and admitting defeat.
On whose instructions were these continuing denials in clear contravention of Intel's corporate policy of openness - made and have they been sacked? Of one thing we can be sure this was not a decision made by an engineer, it was a decision made by an overpaid suit.
And back on that roadmap we saw earlier this week, we find that the i820 and its companion i840 chipset are headed for the gulag. Towards the end of 2001 Intel will put all its eggs in one basket and there will effectively be only one chipset for mid to high end processors the i870. Let's hope this touching faith in the chipset division is not misplaced.
We know the vast majority of Intel employees, and the vast majority of the board, are engineers at heart. When will someone have the courage to stand up to the marketing department and stop them stampeding blindly around like a herd of crazed wildebeest?
The company cannot continue to lurch from disaster to disaster by reacting to outside threats for the simple reason that it patently isn't up to the task. If Intel is to prosper it must hand the reins back to the engineers and launch products when they are ready and not a moment before. It must regain the courage to admit its mistakes. Only then will this great corporation start to regain the respect of the industry.
Intel has damned good people. It used to have damned good management. Can they do something fast before the stockholders sack them? ®
Pete Sherriff contributed to this article
We have changed our recommendation on Intel stock (NYSE: INTC) from 'Hold' to 'Sell'
Sponsored: Hyper-scale data management