by Bob Haavind, Editorial Director, Solid State Technology
While semiconductor capital spending may slump next year — perhaps by 20% — the future remains bright for process equipment companies, and their stocks are actually undervalued, believes Fidelity Investments analyst Peter Wright, speaking at a SEMI breakfast near Boston on Sept. 19.
Wright and Robert Maire, analyst for Needham & Co., did a penetrating analysis of the changes reshaping the chipmaking and equipment industries, but reached somewhat divergent conclusions. Maire suggested that a maturing industry rates lower stock valuations, and that the price/earnings multiples of the ’90s are now history. He believes that consumerization of the industry means that near-term valuations are driven more by macroeconomic forces, and the industry’s stocks may become more like consumer cyclical stocks.
Wright agreed that multiples have been beaten down, but he feels that equipment stocks deserve a premium multiple to reflect the improved business models of toolmakers. He sees a secular improvement in inventory control throughout the supply chain with just-in-time deliveries, better cost controls, shorter delivery cycles, and more cautious spending by fabs. Also, the high cost of developing 300mm tools seriously eroded margins, but these should be much higher over the next few years. The result will be improved earnings and growing free cash flow, much of which tool vendors (particularly the largest ones) have been returning to stockholders though buy-backs and dividends. He further estimated that semiconductor companies are settling into long-term growth of about 9%, with the equipment companies at around 9.5%.
Due to this overall analysis, Fidelity is overweight in semiconductor capital equipment (SCE) stocks by about double the weighting of the sector, at some $8 billion, Wright explained. He also pointed out that unlike semiconductor companies, toolmakers are not highly capital-intensive due to extensive outsourcing, which aids cash flow.
Maire agreed that toolmakers have become much more efficient producers and the industry has matured to some extent, helped by the more conservative, profit-focused spending of chipmakers. But he points out that now that the consumer is king, there is lumpy, seasonal tool buying and hand-to-mouth planning for just-in-time deliveries. For example, 40% of iPods were sold in December last year, and he believes the same may be true this year for the iPhone, with its 8GB of memory. Time-to-ramp capacity is becoming critical as chipmakers respond to market opportunities. More caution by lenders also means there are more capital constraints as the industry matures. Foundries and memory chipmakers are both emphasizing tight capacity management while using the 300mm Prime project to improve tool productivity. He also pointed to the memory chipmakers’ ability to shift between flash and DRAM, with Samsung now shooting for a 30 day turn time.
While the industry has shifted to a demand-driven model, according to Maire, consumer demand has been pushing toward a bell-shaped curve rather than being a leading edge technology driver, which will mean more trailing-edge capacity at much greater utilization rates. He sees much bigger demand for “middle-of-the-road” geometries,” and not as much of a push toward 65nm as we would have seen in the past, because many consumer devices don’t require the most advanced chips.
The big cost disadvantage is killing 200mm memory fabs, so the manufacturers will have to shift to 300mm, and that will keep the memory market alive as a tool buyer for awhile more, Maire said. But it is unclear what will happen to the older fabs, and which companies will be left standing after what he called a “musical chairs” migration — only be a couple of them may be left standing based on who can best afford the 300mm shift.
Maire sees many positives for equipment companies looking forward, but with some qualifications. There is not significant overcapacity and chip inventory levels are in reasonable shape. Cyclical variations have moderated. He does not believe that flash makers will repeat the 1990s DRAM disaster because of widely varied applications, but warned that there is still some concern. Vista and PS3 are “dead on arrival” this year, he said, but he still sees the memory makers continuing to buy tools, and believes there will be single-digit capex growth rates into 2008.
The key for toolmakers will be execution, and Maire cited Lam Research’s financial track record as a leader in this area. Memory exposure is important, and that has helped Lam, he added. Copper may gain a “second wind” as it migrates to memory, but there has been little impact so far. He believes that yield management and metrology companies should lead the industry averages as process complexity increases exponentially. The back end remains the toughest market, while double patterning is pushing back EUV in lithography.
Tool companies are desperately seeking growth into allied industries, but the push into photovoltaics may end up like the flat-panel display market, “only worse,” he said. The problem is that the process machines are simpler and will be commoditized, making it a tough market in the long run, Maire feels.
While Wright of Fidelity agreed that management had become tighter throughout the industry, and that inventories across the whole food chain have been aggressively pushed down for six years, he sees recent signs that inventories may be reaching excessive levels in certain markets, specifically citing DRAMs. He agreed that memory and non-memory cycles have become out of sync, and he expects this to continue into 2008, with higher memory spending offsetting declines for foundries and IDMs. This stronger sustained memory tool buying will prevent the cycle downturn from being severe, he suggested.
But at the same time, the factors that have driven accelerated growth in units over the past five years remain, and he expects more upside surprises ahead. He feels that unit growth has moved onto a higher trend line over the past 8-9 years, and that improved management will drive improving profitability of toolmakers in the future.
From 1990-2003 the equipment makers generated billions in free cash flow, reaching $5.6 billion in 2006, and each year is now better, Wright stated. “I expect the industry to continue to reduce operating expenses as a percent of sales in CY07, as well as improve incremental gross margins and grow free cash flow to record levels,” he concluded.
Wright suggested that new methods must be used to evaluate SCE stocks, since price/book ratios become irrelevant for companies using free cash flow to buy back shares. He believes that forward earnings and free cash flow are more pertinent for projecting upside potential. — B.H.