For all of the excitement about the coming 450mm transition, not everything about the disruption will be be rosy for all sectors of the industry. While there is still some debate on the overall cost savings that will be realized by using 450mm wafers, there’s not much discussion about where the savings really come from. Nothing comes for free.
Beam vs. Non-Beam Tools
The distinction in productivity between semiconductor “beam” and “non-beam” wafer processing tools is quite clear. However, a key, sometimes ignored, part of the cost-benefit discussion, is tool price. To refresh our memory, beam tools are those in which the laws of physics limit the amount of area processed over time; Lithography, Inspection, and Implant are examples. Non-beam tools (e.g. Etch, clean, and thermal processing) generally work on the entire wafer at once, and productivity gains are on the order of the area increase (double).
Most non-beam tool suppliers will be hard-pressed to see a financial reason to be motivated about 450mm. Let’s use a simple example… if a supplier approaches a chipmaker and states they have a new tool that will more than double their current throughput, it would grab their attention—if the new tool costs double, it is likely the interest will wane. However, if it costs, let’s say, only 50% more, it would be a tremendous value for the chipmaker, all else being equal.
Of course, it’s not a good business model to sell fewer tools, higher productivity tools, and overall make less money. How much less? Simple math: If a supplier charges ~1.5 times more for the beam tool, and the customer now needs ~2.3 times fewer tools, the supplier’s revenue drops by 35%. (1.5/2.3 = 0.65). In this case, investing the R&D cash required to develop a 450mm non-beam tool is a tough sell. This is fundamentally the non-beam tool scenario for 450mm. While there’s insight to be gained in this illustration, the actual economics of 450mm equipment are not quite that simple.
Why it’s Not That Simple
When the industry was growing at 20%, 15%, or even 12%, by the few years it took to ramp up the next wafer size (e.g. 100mm to 150mm), the increase in market size and ensuing sales volume more than made up for the non-beam tool supplier’s loss. After about two years, the 15% growth made up for the hit taken by the non-beam tool supplier, post wafer transition. Not only this, the cost reduction allowed the chipmaker to drop prices and drive increased chip demand—everyone benefitted.
Today, however, the markets for electronics are much more saturated than in the early, high-growth period of semiconductors. Now growth is in the single digits, closer to 6% or less. At this rate it takes five years or more for the supplier to make up for the 35% net loss. Fortunately, equipment upgrades, that take place with each node transition, help offset this impact (though much of the equipment is reusable node-to-node)—but how many node transitions will occur in the 450mm era? If 450mm is in high volume for 10nm, there is possibly only one transition remaining (7nm).
Beam tool providers? They have pretty much been given a “pass” with regards to significant productivity increase with 450mm. After all, laws of physics get to take the heat. Likewise, Inspection is one of the first tools needed to qualify every other process and production tool. Therefore, expect companies such as ASML and KLA-Tencor to be able to write their own 450mm tickets while actively engaging chipmakers without nearly as much consternation as will be felt by non-beam tool providers.
What to do?
Raise Tool Prices
Unfortunately, many beam-tool suppliers will be stuck between a rock and a hard place on 450mm. Raising tool prices is an option but competitors hoping to be the last supplier standing and/or startups waiting in the wings, anticipating that 450mm will be the industry disruption that allows market entry, would use increased tool pricing as an advantage; that is, unless the supplier has some intellectual property or service that clearly sets them apart. Conversely, considering the recent supplier consolidation, a supplier sitting on a near-monopoly position may have the market share to raise prices. The problem is that the economic justification for 450mm depends on better cost/performance over 300mm, so there would be tremendous chipmaker pressure on any beam tool supplier that wants to raise prices above what makes 450mm tenable, overall. What is that unacceptable price multiplier? Don’t expect a readily available answer from chipmakers.
Ultimately, what would save everyone is if semiconductor demand were to ramp back up to double-digit growth. There have been many theories proposed on what it would take to make this happen (if even possible). With PC demand dropping, and eventual smart phone and tablet market saturation, new, creative, end-use applications for ICs have to be dreamed up. Perhaps, with low-enough chip pricing, the Internet of Things will help move ICs into places never before considered. Will Google Fiber’s gigabit speed generate new and creative uses of silicon? Do Apple or Samsung have another disruptive technology up their sleeves? Increasing the demand for ICs, through increased cost reduction and new end applications is incredibly hard work, but ultimately is the direction required for all to benefit.