In an earlier column, 28nm – The Last Node of Moore's Law, we pointed out that the change has already happened. It is no longer a matter of forecast or prediction. In this blog, we will start by reviewing some of what has transpired since that column was published. Then we will focus on the ensuing paradigm shift in the semiconductor industry.
The following chart was presented in the IEEE IITC workshop by GlobalFoundries. It illustrates the cost impact of the double patterning required for scaling below 28/22 nm.
In his coverage of Semicon West, Rick Merritt wrote: "Moore's Law has definitely slowed." He quoted Gartner semiconductor analyst Bob Johnson as saying, "No matter what Intel says, Moore's Law is slowing down. Only a few high-volume, high-performance apps can justify 20 nm and beyond."
Soon thereafter, in a Semiconductor Manufacturing & Design blog post about ConFab 2014, Peter Singer quoted Dr. Gary Patton, vice president of semiconductor research and development center at IBM, as saying: "The challenge we're facing now is two-fold. Number one, we're struggling to get that 0.7X linear scaling. It might be about 0.8X. And we're adding a lot more complexity, especially when you adding double and triple patterning."
More recently, in early August, we finaly got more information from Intel about its upcoming 14nm technology node. In our blog Intel vs. Intel, we articulated that Intel's numbers indicate that Moore's Law stopped at the 28/22nm nodes, both in terms of the required bringup time and the cost of the new technology nodes.
It is hard to accept that a trend that has held strong for 50 years, and that kept going many years after multiple predictions of its imminent demise, has really stopped. And it is even harder as we watch the huge effort of bringing up the 14nm and 10nm nodes. Yet it seems that everybody should agree that the semiconductor industry is now going through a paradigm shift and -- for most designs -- 28 nm is, at least for some time, the last node of Moore's Law.
These well-known charts present the reason for that change.
These charts show that design costs increase by more than $100 million from 32 nm to 16 nm. If we assume a die cost of $10 at 32 nm, and if we assume that the traditional cost reduction per node still holds, then we would need a volume of more than 20 million units just to break even. If one also considers the risk associated with such a design, it would actually require more than 100 million units -- or at least $1 billion of market -- for such device to justify the investment. Clearly, very few designs have the market for 100 million units or $1 billion.
The following chart by IBS presents the past trend in design starts per node. As we see, most new designs are still created at the 130nm node, while the node with the fastest rampup is at 65 nm.
Design starts per year; click here for a larger image.
(Source: IBS Dec 2012)
A 2016 forecast from Anysilicon on semiconductor technology nodes is illustrated in the following pie chart.
Yet again, this indicates a very slow shift to more advanced nodes, and the expectation is that -- even in 2016 -- most new designs will still be implemented at the 130nm node. This is clearly a paradigm shift in the industry, which is responding accordingly. Just before Semicon West 2014, we saw the conclusions of the SEMI's World Fab Forecast. This forecast uses a bottom-up methodology, providing high-level summaries and graphs, along with in-depth analyses of capital expenditures, capacities, technology, and products by fab. The following chart illustrates this new paradigm.
The report states:
The cost per wafer has become an increasing concern below the 32nm node. The expected cost reduction benefit of production at smaller nodes is diminishing and is not keeping pace with the scaling benefits in many cases. This has widespread and fundamental implications for an industry long following the cadences of Moore's Law…
These may be contributing factors as to why some volume fabs are exhibiting a lag in beginning production of a new technology node. Now evident quantitatively for the first time, there is evidence of a clear slowdown in volume production scaling of leading technology node transitions. [Emphasis added]
To Page 2