While Intel (Nasdaq: INTC) produced excellent results in the first two quarters of the year, it's been unable to meet customer demand. At the same time, its primary competitor, Advanced Micro Devices (NYSE: AMD), also performed well beyond expectations. As a matter of fact, AMD has executed well enough that you can make the argument that it's actually on track.

As an Intel shareholder, I have to wonder why a company with Intel's resources can't meet demand. To understand why, it helps to look at the accounting-related implications of a company's capital allocation decisions. I apologize up front if there's too much accountant-speak in this report, but it's important context.

We love our Rule Makers because they generate gobs of cash. But generating cash isn't enough. The company must also put its cash to good use. If it can't, it will eventually implode, since poor investment decisions lead to weak results.

Let's take a closer look at why Intel hasn't met demand.

To make more chips, there are a couple of things Intel can do. One is improve its processes to increase chip yield. Intel has done this. It has also enhanced production by converting its microprocessor manufacturing technology from a 0.25-micron process to 0.18-micron, and ultimately to 0.13-micron.

These changes come with a cost. Intel has to invest some cash in research and development. It also has to invest cash in property, plant, and equipment to manufacture these chips. For accounting purposes, the R&D expense reduces net income. This means that, if it doesn't provide positive results, the company will ultimately see its margins fall.

The other way Intel can increase chip production is to build a new chip fabrication plant (fab). That's what I want to focus on here. Since it costs somewhere between $1 billion and $2 billion to build a new fab, this isn't an easy decision. Even with Intel's method of reproducing the same plant anywhere in the world (copy exact) it takes one to two years to build a new fab. As such, a decision to build a new plant today will not result in increased production today. Even worse, if demand levels off by the time the plant is ready to go, Intel's net margins could suffer.

A company's manufacturing costs can be broken down into two components: fixed and variable. Fixed costs don't vary with the level of output. They are costs a company must incur if it expects to keep functioning. They are also costs that must be paid regardless of sales. They include rent, insurance, lighting and heating bills, property taxes, and management salaries. There's one more that I left out. Can you guess what it is?

Depreciation. When Intel builds a plant and fills it up with equipment, the accounting rules require it to write the building and the related equipment off over their useful lives. This is one of the most significant fixed costs Intel incurs. And, if it builds a plant and finds that demand isn't sufficient for it to be fully utilized, its gross margin will fall as the fixed costs are allocated to fewer units. (It's important to remember that, although we calculate gross margin on an overall basis, it's actually realized on a product basis.) When demand is high, one of the reasons gross margin increases is that fixed costs per unit are lower. If an excess of capital is allocated to building new manufacturing facilities that can't be fully utilized, the end result is that margins will be adversely affected and the company's long-term financial picture is weakened.

Variable costs, which move in close relation to changes in sales, are tied to the number of units sold. They include raw materials and direct labor costs. If Intel builds a new fab and it doesn't run at full capacity, then it doesn't incur additional variable costs.

This is an important distinction. Variable costs represent the direct input costs related to each unit produced. Fixed costs represent static costs that are allocated to each unit produced.

Right around the time that I first purchased shares of Intel (about five years ago), there was much debate among the Wise about whether Intel was overbuilding. In retrospect, it seems that it underbuilt. To some extent, those decisions led to Intel's inability to meet customer demand today.

This raises interesting questions. If both Intel and AMD were able to meet customer demand on their own, who'd win out? Would Gateway (NYSE: GTW) have signed up with AMD for Athlon chips if Intel had been able to meet its capacity needs? Would Intel have had idle plants that hurt gross margins? These are tough choices and I don't have the answers, but they are interesting issues to contemplate.

This discussion touches on a couple of other issues. One of the reasons that Rule Makers continue to thrive in the market is their dominant cash position. Another is the importance of topflight management teams that can make decisions and map out a successful approach for the future. Dominant Rule Makers have a big edge over the competition in capital allocation decisions. They have more cash to spend without crippling their business. Particularly in capital-intensive businesses, this makes it easier for them to maintain their lead.

After giving these issues further thought, I can better understand why Intel doesn't have sufficient manufacturing capacity. But, I also realize that this situation gave AMD an opportunity to succeed that might not have been there if Intel had additional capacity. So far, AMD has taken advantage of the opportunities. Its gross margins are up and it's paying down debt. So far, however, it's been unable to make serious inroads into corporate America, where the real money in this business is made.

Related Links:

  • Intel Looks Strong Inside, Rule Maker Portfolio, 04/19/00
  • Intel Looks Marvelous, Rule Maker Portfolio, 07/19/00