Thursday, May 03, 2007

Semiconductor Cheatsheet

I cannot understate the value of creating cheatsheets for every sector that one wants to position themselves in Venture Capital. It is not only important that you understand the investments in the sector, but you also need to have an opinion on the state of the industry. I will take this opportunity to put together one such cheatsheet for whatever it is worth.

Since my positioning has always been Semiconductor/Mobile Services/Telecom/India, it is obvious that I get asked a lot about the state of Semiconductor industry. Now even though it does look quite morbid at first glance, there is a lot going on in this sector. One has just has to look a bit harder. Below is my cheat sheet on the Semiconductor sector.

Key Characteristics
• Can be capital intensive and will give less than lofty returns
• Two primary ways of making money- either find Chinese based manufacturers or find extremely capital efficient US companies
• Investors in this sector require significant technical knowledge
• Moore’s law is nearing physical limits and so there is a search for alternatives. These can be Nanotechnology, Bio-electronics, Bio-sensors or something else
• Out of these technologies, Nanotech has the brightest future. Even though single nano-tube circuits are working, there is no interconnect technology yet to scale it. At this point, that technology is 10 to 15 years out. (Nantero claims to solve this, however I have not done any technical diligence on them. Last heard, Draper invested in them and this is always a good sign)
• Out of the currently viable technologies, 3D assemblies, parallelisms and optical interconnects can keep the Silicon CMOS viable (22nm)

Bottomline: In the end, it is important to know the technical obstacles in this industry, and then keep an eye out for those startups that claim to advance these limits.

Current Trends in the sector
• Rise of fabless semiconductor companies
• Increase in Design jobs worldwide
• Customer base is shifting to Asia
• Not surprisingly, manufacturing is also shifting to Asia
• Huge R&D spend on future technologies like Nanotechnologies
• Wireless devices, Video, consumer electronics driving semiconductor demand

What about those Fabless Investments?
Most fables companies take in about 30M in 2.5 years to get the chip out. To get 50% margins, you are looking at revenues of around $60M. To get this for high priced chips, you need volumes of 2M x $30 and for low priced chips, you need 12M x $5. In either case, this involves a significant sales effort. So firms have to either decrease cost upfront or delay getting cash till the cost of capital declines.

Bottomline: Things to look for - Not an incremental technology like a cheaper chip or a bit faster chip, but unique IP and most importantly, a great team. Given the requirement for capital efficiency, it is very important one backs seasoned entrepeneurs in this sector.

Finally, What are the two biggest barriers in this sector?
• Exponentially soaring design costs (costs are around $50-75M for 45-65nm nodes)
• Low chance of an IPO

This cheatsheet coupled with an investment thesis can be a staging point for a good discussion on the semiconductor market. At the end of the day, creating such quick notes for each sector that I am positioning myself in has helped me immensely.

No comments: