Normally, that would probably be it for the year, as sales would fall off from such lofty heights and the industry struggle to digest these orders into leading-edge fabs.

Not so, according to Eric, as there is more to come in 2007!

He expects ASML to grow further this year, beating the figures listed above. Also, it must be said, his justifications were not picked out of thin air - rather there was an element of rational logic behind them, though I hesitate at saying that I was totally convinced!

Meurice thinks that should the IC unit sales forecasts come in at the average of positions being held by the market forecasters, chip manufacturers will need to continue to add capacity this year. This, in turn, should see approximately 3 percent CapEx growth for the year, pushing sales above 2006 levels.

However, he also sees the lithography market in particular outgrowing the average at approximately 5 percent for 2007. So this growth should be a little bit extra in general terms for the litho boys.

But he went further than that by highlighting that ASML's own strength in technology and market share gains in ArF immersion lithography would see even better sales in 2007 as more customers adopt ASML's offerings in that field than they do his two competitors (Nikon & Canon).

The customers doing the buying not only include the major DRAM manufacturers from Korea and Taiwan but improved buying from foundries will also help ASML as these are the ones spending in 2007 and are long-standing ASML clients. Coupled with 7 new customer wins in 2006 and potential for new business from these manufacturers, ASML sales look like surpassing 2006 figures with ease!

To bring this aspect down to earth slightly, Eric chucked-in a final affidavit by noting that due to the incredible value vested in its current backlog of tool orders, if none of the above actually came true, ASML would still have a very good year!

A slam-dunk!

I had written in October last year that ASML was doing so good then that it must be near quarterly shipping capacities. The backlog has definitely grown and grown for the most technically advanced tools (immersion). Shipments for 1Q07 were put at a slightly lower level than what was achieved in 4Q06 so in my opinion ASML is maxed-out right now.

But just when I thought I might have a point or two to perhaps dampen Eric's enthusiasm, ASML announces in the conference call that it will spend approximately €150 million this year on expanding its assembly and test facility at its HQ in The Netherlands. Damn and blast it!

So if there were a weak link in the Dutch armor, what on earth could it be?

Excluding the obvious macro semiconductor economic aspects that may or may not pan out as stated above, after a little pause for thought I came up with the following list.

• Memory over-capacity, especially in NAND.
• Soft demand at major foundries for most of the year.
• Toshiba Fab 3 & Fab 4
• PSC/Elpida JV fabs in Taiwan

With memory over-capacity looming, especially in NAND, I think I have a pretty good argument. With over 50 percent of ASML's sales going into memory fabs, the over-capacity scenario would hit ASML hard. Cancellations, push-outs and even order denials would soon appear and ruin Eric's rose-tinted view of the semiconductor world, right now!

Not to be, says he!

According to the gospel of St. Eric, we are not going to see over-capacity in the memory market, as more of the 300mm fabs out there are now classified as ‘hybrid.' What that means is that the larger memory manufacturers will use their large scale facilities to produce DRAM when DRAM is hot and NAND when NAND is hot. At some point in the curve the fears of real over-capacity will be diminished due to shifting device types being manufactured.

That has certainly happened in 2006 but will it be so good in 2007 when IM Flash Technologies and Toshiba are hell-bent on NAND capacity expansions and have no interest in adding DRAM to the mix?

Are Nanya and Inotera going to back-track DRAM ramps at two new fabs and switch to NAND production when they have never produced NAND devices?

Though I cannot detail here the dynamics currently at play at 300mm memory fabs due to space/time constraints as well as the fact we have just printed a detailed review of this in the latest print edition #32 of Semiconductor Fabtech, let me just say that I am not convinced NAND wont be in trouble this year. If so, that's one weak link in Eric's arguments.

Foundries are fickle things by the nature of their business. Capital spending didn't really happen last year and doesn't look like coming back till 2H07. TSMC is spending, but less than the others. Not sure yet whether foundries will help litho sales that much unless major demand appears by the 2Q07 timeframe.

The other two weak links are basically Nikon customers, with Toshiba perhaps being the dark horse here. The reason being that although ASML has won some new business at Toshiba the majority of orders will go Nikon's way in 2007. The secret part of this is that Toshiba is being very aggressive with immersion lithography as well as 300mm ramp rates. The scale of Fab 4, currently under construction, is so huge it will be the largest fab ever built. More on this in the latest #32 300mm activity report.

All I can say here is that that fab will have a huge number of Nikon immersion tools eventually, as will the PSC/Elpida 3 fab complex in Taiwan. Though not enough to overtake ASML, these alone may close the current gap significantly. Before you jump on this point, yes, they are both memory players and could suffer in the same scenario as ASML, but the point is that at some stage immersion orders from these two customers shouldn't be underestimated.

Even with my cautionary tales and attempts to dampen Eric's claims for 2007, there would seem to be just cause in thinking - more of the same? Apparently so!