We do not spend much time looking at materials for semiconductors. This is an important but highly specialized area that does not often have an impact on the downstream part of the industry where we usually swim. However, there are some important developments percolating up from here and we thought they merit a bit of attention. Most specifically we are talking about Gallium Nitride (GaN) and Silicon Carbide (SiC).
Most of the semis we deal with are built on “standard silicon”, which is more or less what it sounds like – silicon refined into pure wafers that conduct electricity just the right amount. But silicon is not the only material we can use, and for years we have used a whole array of other materials like GaAs and SiGe, to name just two. For pure digital functions standard silicon is the right material, but for many other uses other materials have advantages. And every decade or so, we see the introduction of new materials into the mix. With semiconductors now being manufactured into pretty much everything and the world moving towards more electrification, we are poised to see significant growth in some of these.
GaN and SiC are rising to prominence now for two main reasons. First, all those new electric things are becoming large enough end markets to merit investment in better materials for them – this is the economic angle. Separately, the industry has made some big improvements in making them, making them commercially viable and able to scale up – this is the manufacturing angle.
We are not going to spend much time on the manufacturing side as it is fairly straightforward. Better manufacturing techniques mean that these materials can now be processed in much higher volumes.
On the other hand, the economic side is starting to look much more interesting. The growth in electric cars, large scale energy grids, solar cells, lighting systems and growing “smart” industrial systems all run workloads at very different power levels than more common electronics and will need these new materials.
And here we have to touch (lightly) on physics. GaN and SiC are what are known as wide bandgap materials. They hold electrical charges differently than other materials which essentially allow them to handle higher power levels more efficiently than standard silicon. We do not think of it much but it is something of a miracle that all our day to day electronic devices can all be powered by the same wall socket. Yes, different countries have different voltages (and sockets) but modern electronics deal with this so well we can largely ignore it. In fact we are more likely to complain about the little power cord brick that makes this possible. But many other electric things need to handle higher power loads.
We are greatly oversimplifying this, so here is a better primer. We also found the diagram below on Infineon’s web site, which we think illustrates all this nicely. The grey arc shows traditional silicon, the orange SiC and the fuchsia GaN. The axes show the trade-offs between power (watts) and frequency (Hz). Watts matter for things like trains and energy systems, Hz matters for things that require a lot of fast switching (think refresh rates on monitors). The main takeaway from this chart is that many applications are pushing the limits of what standard silicon can offer and the new materials extend the frontier considerably. It also highlights the different use cases for SiC and GaN. Finally, we should point out that while the circles are roughly accurate, the location of all those applications are a bit subjective, and so other versions of this diagram will show more or less in the SiC or GaN circles depending on who is telling the story.
“”

Does this mean it is time to run out and buy GaN and SiC stocks? Not quite. These transitions take a long time. A whole new supply chain has to emerge – someone to make the materials, refine them, create wafers and then fab them. Many of the large analog chip makers as well as emerging ones in China are already fairly far down this path. Companies like MACOM have been talking about GAN for a decade, and we can still argue how much of this matters to their stock price.
That being said, the demand for GaN and SiC is clearly starting to pick up. We are seeing more articles like this one all the time. The demand is out there and every reasonable forcecast demonstrates that the kinds of devices that need these materials are poised for very strong growth.
(And just to be extremely clear we are not talking about the stocks with tickers GAN and SIC because those are totally unrelated.)