Hi Jacob,
Thanks for posting.
The Gartner forecast is for 2010 through 2014. I am sure it uses a combination of top down and bottom up forecasting with a caveat that it can not and does not account for unforeseen macroeconomic and geopolitical events. It is a qualitative model expressed in quantitative terms. Models which use historic data to predict are also qualitative. 2015 is interesting for what it might imply about the qualitative model.
Many semiconductor companies have trouble forecasting more than the next quarter. Semi-equip companies have greater visibility because they have and see longer lead times both for equipment and empty fabs. Unfortunately, the guys building the empty fabs and ordering the equipment are the semiconductor companies with the short lead times. I think the whole industry is doing or using Gartner type analysis to augment their limited visibility.
On the semi-equip cycle, you are using factual evidence to draw a conclusion. The conclusion is based on fact, but is not fact. I disagree with your conclusion, but it is an important conclusion and needs to be addressed.
First, there is the all important definition of semi-equip cycles. Semi-equips are capital equipment and sales of capital equipment fall during macro economic down turns. This variation of the cycle will always be with us, always be true. There has been another variation of the semi-equip cycle, one that is generated by the internal dynamics of supply and demand within the industry and which has occurred during macro economic expansions. This is the cycle this thread is suggesting will be different going forward.
The data you referred to is associated with the "Great Recession." It is not surprising that the most severe recession since the Great Depression produced "the lowest trough" and the "greatest increase," but this data doesn't say anything about cycles generated by internal industry dynamics.
Your second point is important, also. Semi sales are at a record, but semi-equip bookings are not close to a record. There is an associated piece of data that is also important. Semi-equip revenue as a percentage of semiconductor revenue has fallen in recent years. Again, I disagree with your characterization, "bad news."
First, one would expect a lag between record semiconductor revenue and record semi-equip orders. Particularly, after a severe downturn and with macro economic problems still in the news, it is only prudent to run fabs at high rates, limit supply to maintain or lift prices, and invest cautiously.
If semiconductor companies have "gained bargaining power," it would be reflected in the semi-equips margins, particularly gross margins. I follow AMAT, ASML, KLAC, and NVLS closely. I am hearing that they will produce record margins at various recovery revenue levels. Also, I have industry forecasts that say record revenues will be attained by AMAT in 2011, 2012, ASML, 2011, 2012, KLAC, 2012, and NVLS 2011, 2012. The years are the companies' fiscal years.
The issue of falling semi-equip capital intensity has been more persistent and needs to be addressed. There are a number of factors, here. Factors which increase semiconductor revenue and which decrease semi-equip revenue.
300mm - The move to 300mm, 12 inch wafers, had a significant impact. Semi-equipment for 300mm was not much more expensive than 200mm equipment and throughput rates were maintained. The revenue per wafer was markedly increased.
Lithography - 193 nanometer lithography has been the tool of record for many nodes on the Moore's Law road map. Now, an immersion version handles critical layers. This has been the longest period without a change in the frequency of the light source. EUV will bring the first big change in a while.
Rational Procurement - At one time, a fab was built and filled at one time. This produced an equipment bonanza for the semi-equips, but over supply for semiconductor producers. Much of the equipment was not used at high rates and their output sold at low oversupply prices. A very bad way to get high capital intensity. More recently, fabs were equipped, one line at a time. This meant that equipment would be used at high rates, supply and demand would stay more closely in balance, and good prices would be received from their output. This produces lower capital intensity, but it is more profitable for the semiconductor makers. This is also better for semi-equip makers because rational procurement minimizes industry generated cycles and allows the semi-equips to maintain a more efficient production model, one with much fewer changes in production capacity.
Efficient equipment and processes - There is competition in the semi-equip industry. One important competitive factor is the cost associated with a process step in wafer fabrication. Another is the throughput at each process step. Equipment vendors continually work on improving these, but don't raise prices when they succeed. They get the business. These improvements increase semiconductor revenue and decrease capital intensity.
Again, this might all be "bad news," but the 4 semi-equips I follow are expecting record revenues and record margins. |
