by Bob Haavind, Editorial Director, Solid State Technology
An insightful analysis of what chipmakers need to do to thrive in a period of “profitless prosperity” by Steve Newberry, president and CEO of Lam Research, set the stage for a spirited Wall Street panel at the Industry Strategy Symposium at Half Moon Bay, CA (Jan. 16).
Driven by consumer demand, unit growth for semiconductors has been stronger than in the past dozen years, according to Newberry, yet revenue growth of chipmakers has been lagging due to low average selling prices (ASP). He likened the current period to what was termed a time of “profitless prosperity” by W.H. Krome George, CEO of Alcoa, in October 1973, in reference to the aluminum market of that era.
The current period is marked by accelerating IC growth at a time of declining profitability. No matter how much unit shipments rise, Newberry commented, companies “can’t be prosperous if prices are declining faster than costs.” His analysis shows average profits for the semiconductor industry to be about 12% in 2007, which superficially appears to be providing enough cash to finance needed growth, but he questions whether the average presents a true picture. In fact, he went on to explain, a sector-by-sector analysis shows a much bleaker picture, with the outlook dim for almost all the players in every market sector — 15 of the top 40 companies are losing money and 23 are making less profit than needed to stay in the business, Newberry’s analysis showed.
The memory market, where Taiwan and Korea have been jockeying hard for market share, looks particularly tough. The expected CAGR unit growth from 2007-2010 is about 25% (vs. 50% growth in 2007), which will take some $83 billion for equipment out of $128 billion in total capital spending, Newberry said. Assuming a 25% tax rate (some vendors pay more) and depreciation of 32%, he concludes that a 17% operating profit will be needed to finance future growth. No players in the memory business are close to that, he said. “In 2007 NAND flash and DRAM makers were not making enough money in a 50% unit growth year,” he concluded. Earlier in the ISS, Bill McClean of IC Insights showed a chart (see Fig, below) tracking unit ASPs for DRAM from nearly $7 at the beginning of 2007 to below a dollar by the end of the year: a precipitous decline!
ASP trend in 2007 for 512Mb DDR2 667Mbps DRAM chips. (Source: IC Insights, DRAM eXchange, Elpida)
The trouble is overinvesting to create excess capacity, trying to force-feed a much larger industry than exists, Newberry commented. Too many vendors are trying to capture the same market share through rapid supply line ramps that don’t allow much differentiation, he said. The only sectors that he found to be financially healthy were analog and fabless, which have low capital investment requirements.
In the logic sector, Newberry believes that no integrated device manufacturer (IDM) can achieve enough volume to effectively compete with the economy of scale of the foundries. Even Texas Instruments, the only IDM with profit over 10%, is moving from a fab-lite to fab-liter strategy, he commented.
Even in the foundry sector, all the profits in the past five years were made by TSMC, while UMC, Chartered, and SMIC are not generating sufficient operating profits to sustain growth, he concluded. How is TSMC able to compete effectively, since it does not appear to have any cost advantage versus its competitors? While there is some economy of scale involved, Newberry believes there is a more important factor — TSMC is able to command a price/wafer premium by offering superior value to its customers in terms of libraries, extra services, design aid, etc.
Newberry pointed out that the major effort of most chipmakers was to solve the profitability problem through cutting costs — but a quick analysis showed that even if toolmakers cut their costs enough so that the entire process tool industry made zero profit, it would only offer chipmakers savings of perhaps $5.7 billion, when they need about $11.2 billion to close the profit gap, he believes.
Thus, Newberry suggested that the chipmakers’ traditional business models with strategies of the past would not prove sustainable. Great improvements have been made in leadtimes, inventory control, productivity, yield, and efficiency, but these have not solved the problem. They must do better end-demand analysis and improve their demand/supply forecasts — but even more, he said, they must focus on better value creation for customers, differentiating themselves while also becoming leaders in efficiency. — B.H.
Next week’s WaferNEWS will continue with the lively debate that followed Newberry’s comments among four top industry investment analysts, led by moderator Jim Bagley, chairman of Lam Research.