Ask About Computers Blog

There’s a new CFO in town at Intel. The preceding CFO has moved on to become Chief Operating Officer. The former Assistant CFO is now the CFO. The new Chief Financial Officer received baptism by fire in an earnings conference Tuesday (Jan. 15, ‘08).

You would have thought it was a good quarter. Intel shipped a record number of server CPUs. Mobile CPUs were very strong, with processors and chipsets setting record units. And desktop quad-core shipments were up 40%. The CEO did mention that flash “had a challenging quarter” (p. 2).

Truth be told, stocks as a whole had begun to tank a couple of weeks earlier. In such an environment of fear, it needed but a spark to keep things tumbling.

Sales at Intel, while setting records, nevertheless failed to live up to expectations. Sales in turn affected profit. Also troubling, Intel guided below Wall Street’s expectations for the first quarter. People scanning the headlines read into these figures, the sky is falling.

Intel, however, insisted that as of yet it saw no signs of economic slowdown. Many at the time chose not to believe Intel management. Hindsight is 20/20, it is said. Today we know that Intel was not lying. Other IT companies, such as AMD and Seagate, have reported much the same thing. There’s no slowdown yet affecting these companies.

It was Intel’s NAND flash memory business that had caused the sales miss, not microprocessors or anything else. This is not to say that a worldwide recession won’t happen, and if it does Intel will be affected. NAND sales, however, are down worldwide because of oversupply.

Almost right off the bat, the Intel CFO and analysts were at odds over gross profit margins. For the record, Intel improved its gross margins. The CFO, however, refused to break down estimates for margins for upcoming Q2 through Q4 quarters. He gave an estimate for Q1 and for the year. He could see two points of good news affecting margins for 2008. Beyond that uncertainty.

The unknown variable was competition. Pricing is “a function of the strength of our product roadmap relative to competition” (p. 6).

AMD, after a slow start in ‘07, will continue to ramp Quad-core Opteron throughout 2008. If AMD can get the clocks up of Barcelona, it should give meaningful competition. Depending on the competitive environment, Intel’s gross margin “could go higher, it could go lower” (p. 6).

The analysts for their part didn’t want a breakdown of margins quarter by quarter. They just wanted to know what the forces of evil were weighing down gross margins. The CFO responded, “I don’t have negatives that are hidden in that margin forecast” (p. 5).

Why then is the estimate for the year so low, one analyst asked? The CFO: “It is the gross margin forecast”.

Stalemate. You could hear the frustration in the voices.

The analysts got the last laugh when many went home and “cut estimates and price targets on the stock” (Tech Trader Daily).

Read the transcript here.

It may well be that, after a big drop, “the price of NAND flash is showing signs of stabilizing”.

However, the problem this past year hasn’t been demand. The demand’s been there. The problem has been supply. There’s been an oversupply of chips on the market for the larger part of 2007.

This would suggest that times will not get better for memory makers until memory manufacturers trim capacity and burn off inventory. Also on the more distant horizon there’s solid-state drives which should drive demand like there’s no tommorrow once they become popular.

Memory makers made great preparations for Windows Vista. They started to manufacture memory chips on larger 300-mm size wafers. Vista didn’t take off quite like expected, and now there’s a glut on the market of memory. It’s a good thing demand has stayed strong or things would be much worse.

In the meantime, chipmakers are cutting back on their capex for chipmaking equipment. What is bad for chipmaking equipment companies should drive the next up cycle in the memory business. At least one memory company is even rumored to be trimming production.

Read more here.

One of the things that makes semiconductors interesting is the on-going collaboration and potential consolidation taking place in the industry. The cost of manufacturing chips is so great that companies are having to pool their resources to get the job done.

A good example of this is the newly announced collaboration between memory maker Qimonda and Macronix. A glaring omission from Qimonda’s portfolio is flash memory. So Macronix shall supply the flash technology, while Qimonda supplies state-of-the-art 300-mm silicon wafer technology.

Qimonda + Macronix may make NAND flash, since this is the memory of the future.

Read more here.