On Dec. 22, 2009, I wrote an article on TheStreet.com describing how Intel (INTC) - Get Intel Corporation Report was finally making money on the Atom processor with its new (at the time) Atom N450. The basis of my argument was a press release from the company that "the new single core chip, called Atom N450, is about 60% smaller than present Atom processors and will require 20% less power."
It has come to my attention after searching Intel's Web site that the N450 has a die size of 66 sq. mm, much bigger, NOT 60% smaller, than its predecessor, the N230, with a die size of 26 sq. mm.
Why the smoke and mirrors? Intel has been on the defensive ever since
The Information Network
in January 2009 questioned whether Intel's slashed revenue of about $1 billion (for the fourth quarter of 2008 and first-quarter 2009) was because of its misjudging the success of the netbook and its Atom processor.
Paul S. Otellini, president, chief executive officer, and director of Intel, made the following statement during Intel's second-quarter 2009 (July 14) earnings call.
"Atom revenue grew a very strong 65% this quarter, reflecting a bit of a snap-back after the inventory correction we saw in late Q4 and in Q1," which suggests to me that Intel conceded to
The Information Network's
Based on new size information, I need to recalculate Intel's revenue for the netbook market. A 300mm wafer will make 1,071 of the new N450 Atoms vs. 2,716 of the old N230. Based on a price of $64 for the N450 and $29 for the N230, each 300mm wafer will generate $68,500 for the N450 and $78,800 for the N230.
According to a report from
, 31.1 million netbooks are forecast to be made in 2010. To generate that many notebooks, Intel would need to process 11,400 wafers of the N230 Atom and 29,000 wafers of the N450 Atom. Based on the above revenues per wafer, Intel would have made $902 million in revenues for the N230 and $1.990 billion in revenues for the N450: $1 billion more with the new N450.
That's not all profits, as processing 29,000 N450 wafers vs. only 11,400 N230 wafers is not cheap. I went to the trouble of calculating prices based on wafers because production costs in Intel's fab are based on wafers. You would get the same revenue values if you just multiplied the number of netbooks times the price of each chip.
Based on some calculations I've seen, and keeping in mind that manufacturing costs are highly company dependent, processing a 300mm wafer at 45nanometer costs about $5,000. The cost to Intel to manufacture the 11,400 N230 wafers would be $57 million, while the cost to manufacture the 29,000 N450 wafers would be $145 million. Thus, net profit would be $845 million for Intel if all the 31.1 million netbooks were made using the N230 Atom vs. $1.845 billion using the N450, still about $1 billion difference.
The above calculations are based on Intel holding a 100% share of the netbook market. However, in 2010,
projects Intel will hold only a 70% share, losing out to the
processor, which will eventually gain a 55% market share of the 96.0 million netbooks sold in 2012.
With a 70% share, Intel stands to make $1.3 billion (not the previously estimated $1.6 billion) in profit on netbook sales in 2010 because of the N450 Atom compared to only $590 million had they continued to use the N230.
Robert N. Castellano, Ph.D, is President of The Information Network, a leading consulting and market-research firm for the semiconductor, LCD, HDD and solar industries. Castellano is internationally recognized as one of the leading experts in these areas. He has nearly 25 years of expertise as an industry analyst. Castellano has provided insight on emerging technologies to many business and technical publications, including Business 2.0, BusinessWeek, The Economist, Forbes, Investor's Business Daily, Los Angeles Times Magazine, The New York Times, USA Today and The Wall Street Journal. He is a frequent speaker at conferences and corporate events. He has over 10 years' experience in the field of wafer fabrication at AT&T Bell Laboratories and Stanford University before founding The Information Network in 1985. He has been editor of the peer-reviewed Journal of Active and Passive Electronic Devices since 1985. He is author of the book "Technology Trends in VLSI Manufacturing," published by Gordon and Breach. His book "Solar Cell Processing" was published in 2009 by Old City Publishing. He received his Ph.D. in solid state chemistry from Oxford University.