Intel's obsession with marchitecture is hiding real problems Part II Charlie follows Chipzilla's spoorBy Charlie Demerjian: Tuesday 04 April 2006, 07:01 This is part two of the article which started hereINTEL IS DISTRACTED and is spending an inordinate amount of money on end user marketing. It is trying to re-create the Centrino magic with media centres and business desktops. There are two problems here, product and money.The product part is simple. Centrino was a smashing but expensive program based on an inherently fabulous product, the Pentium M chip. The other two programmes, Viiv and Pro are not based on a category killer product. Was Centrino successful because it was a branding coup, or because it was something that people already wanted?Intel had a great product at the right time, and that lead to success. If the branding had anything to do with it, AMD would not be gaining as much share in the segment as it is now. If you have a good product and tell people about it, they will buy. If you have crap and tell people about it, you may sell a few, but it won't have legs.Intel seems to have lost its focus, from messes like Itanium, LCOS and networking chips that drain engineers, management mindshare, and resources from things that matter, X86 CPUs. Intel seems to be putting non-infinite engineering resources in the wrong places with catastrophic long term effects.These distractions are bad enough, but it doesn't stop there. Every time it takes people off projects that will make the foundation for the future and put them on crash checkbox engineering programs, it eats its own future all that much more. The engineers at Intel see this, Pat Gelsinger gave a stirring performance per watt speech years before it mattered, he saw it, and presumably it wasn't a fluke. It wasn't acted upon, and the Pentium 4 mess happened instead. How much did that cost? Why was Intel focused on marketing instead of technology? Good products sell themselves, ask AMD.Management, ManagementThis brings us to management and we believe Intel management is simply not doing its job. It stays busy patching up PR problems instead of addressing what really matters. By all accounts it is out of touch, and fostering an atmosphere where it pays to keep your head down and shut up.Intel does not seem aware of how much of its lunch AMD is eating, and that is very bad. The other one is that people at Intel have hunkered down to 'let's not disseminate bad news' mode, and that creates a toxic environment that no corporation can live with for long. Management is out of touch and not getting the information it needs. The Intel of old would never have let this happen, or at least never let it happen publicly.Then there is Intel's much vaunted planning capability, one that is falling flat. As was said earlier, they are awash in parts, and from the looks of it, it is getting worse by the day. When you have a product manufacturing cycle that if you turn it on a dime and execute perfectly, you won't see a change for three months, what happens when management does not see a problem until they read it on the Inq?If you have your fabs churning away at a level to serve 83% of the market and you are at 79%, there are a lot of chips that you have to get rid of somehow. Intel is caught in a cycle of executive myopia that is feeding on itself. It doesn't see the problem until it is mainstream headline material, and by then, the current quarter is lost.To make matters worse, Intel has cut prices using rebates and desperately hopes no one notices. Chips have the shelf-life of seafood in a tropical climate, so you have to do something with them quickly, but what? If you give distributors kickbacks, you cut your profit by large chunks. Dropping prices has a similar effect on the top line. Pentium 4 price drops are very hard to undo, and the effects of a cut can take several quarters to filter out.Celeron, ever onSo, what has happened? Intel has fused chips down to stupid levels, and pick a part that does 'not exist', pick a part number that fits in, and hope no one notices. It then sells these Celeron4 2.4/533s at a blow out price and floods the market with them. This protects the prices of even your lowest end parts, and moves inventory. Intel has been doing a lot of this lately, way too much for comfort. There are a lot of these parts floating around, and eventually, that will make the real Pentium 4s harder to move.The problem is that Intel needs to attack AMD at the top end, but has no ability to do so in a comprehensive way. To do that they need a solid platform, and they don't have one. Management is covering massive mistakes made in the past with sleight of hand while the core problems are not being addressed. In the end, we are left with a company that is not addressing the problems that it needs to. Intel knows what the problem is, knows how to fix things, and has all the resources to do so. But instead of doing the obvious, it is spending the capital on short term stopgaps and PR spin. That money is flat out wasted and allows AMD to widen the technological gap at its leisure. If this was a single problem, a missed product, or a marketing misstep, that is understandable, because everyone makes mistakes. What we are seeing at Intel is a complete inability to grasp, comprehend, and engineer an answer. It has now been years since it first admitted there was a problem, and that admission was far too late to begin with.Since then, it has made one misstep after the other, and there is absolutely no sign that things will improve until 2009 at the earliest. Until then, AMD has something of a free ride. There is only one thing to do and that is to fix the management because it seems the current team is unable to do what is necessary to turn things around. How much more bleeding will happen until the board wakes up, or more likely the stockholders revolt? Change needs to happen soon, and until it does, the elements necessary to start the rebuilding are incapable of getting off the ground. It has had several years and two administrations to make the change and it is time for a new approach.
