Sunday, May 25, 2008

Momentum for MEMS

The MIT Club of Northern California's Semiconductor Entrepreneurship series hosted a keynote address on "MEMS and Startups" by Dr. Kurt Petersen, on May 15th. MEMS are microelectromechanical systems and at nano-scales classified as nanotechnology also. Dr. Petersen is considered one of the founding fathers of MEMS and has also co-founded four startups, Transensory Devices, NovaSensor, Cepheid and SiTime. Dr. Petersen's 1982 review paper "Silicon as a Mechanical Material" is considered one of the seminal papers in the field. You can find his bio summary on the event link. The talk was fascinating in that it covered the emergence and growth of a new industry and the fate of four startups in the space, a story often repeated with some variety in Silicon Valley.

MEMS devices have grown by an order of magnitude every decade from the 1970s. In the '70s the applications were mostly industrial, followed by medical in the '80s, automotive in the '90s and consumer in 2000 and beyond. the volumes have grown from the millions in the '80s to billions currently. Early applications in the '70s were pressure sensors and thermal print heads. Today they cover areas such as pressure sensors, accelerometers, digital light projectors, gyroscopes, oscillators, inkjet print heads, microphones, microfluidics for biotech and many others. In 2006 MEMS was estimated as a $5.6B industry with TI, HP, Canon and Bosch leading the pack with revenues of $905M, $500M, $427M and $374M respectively.

MEMS is everywhere in automobiles with all kinds of sensors, and it is expected that every cell phone will have 6 or more MEMS chips by 2013. Early MEMS devices were built in speciality foundries, but today big foundries like TSMC have announced plans for full scale entry into MEMS . MEMS is rapidly moving to state of the art 6" and 8" wafers. When devices such as the iPhone and Wii use MEMS, clearly there is enough volume to attract the bigger players. With the big foundries entering MEMS, the possibilities for fabless MEMS companies also increase. In fact, Dr. Petersen's SiTime is one such company. He recalled the early '80 and early '90s when they had to beg foundries to build MEMS devices.

Dr. Petersen talked about his experiences with his startups. The first, Transensory Devices founded in 1982 is now part of Measurement Specialties. The second, NovaSensor, founded in 1985 along with Janusz Bryzek and Joe Mallon was funded by Schlumberger with $2.5M in equity and $2.5 M in R&D. The company shipped second source pressure sensors within 7 months of funding and became a first source supplier because of quality and delivery. The company was sold to Lucas Industries in 1990 and acquired by GE sensing in 2004. Today NovaSensor has over $100M in sales. Dr. Petersen described milestones in the company's history such as getting the first borrowed fab, building their own fab after a year and a half, getting sued and going through two layoffs.

Cepheid was founded in 1996 and self funded for 18 months. From 1997 to 2000 they received three rounds of funding for a total of approximately $24M from angels and CampVentures and went public in June 2000. They shipped first product in 2000. Cepheid's mission is to change the way DNA testing is performed using microfluidics and PCR. All US mail is screened for anthrax in Cepheid systems. Today Cepheid has over 500 employees and a market cap of close to $1.4B and derives most of its revenue from the human diagnostics market, but is not yet profitable. The revenue from the USPS is about $40M/year and CPHD is one of Silicon Valley's top 150 companies with a revenue of $130M in 2007. Their GeneXpert has 3 FDA approved human tests and many more under development.

SiTime was founded in 2004 and received three rounds of funding totaling about $43.5M from NEA, Greylock, CampVentures and Bosch. SiTime's mission is to make the first and best MEMS based oscillators in the market. They seek to beat the quartz oscillator cost model of $0.50/oscillator. SiTime shipped first product in volume 33 months after start and is currently shipping over 100k units/week. SiTime is one of the early fabless MEMS companies. Quartz oscillators are used in a variety of devices from watches at the low end to consumer electronics and very low jitter high end devices. MEMS oscillators are initially targeted at the mid-range consumer electronics segment and will penetrate the other markets over time. SiTime is still a private company.

Based on experiences with his startups, Dr. Petersen talked about key startup requirements as being an experienced team which could weather adversity, a market over $200M and advanced technology with IP protection which was inexpensive and met market needs. However, the reality of startups is that they are never perfect, never easy and make mistakes. They also have bad hires, bad times, conflicts due to multiple type A personalities, are frustrating and have poor initial products. However, if the founding concept addresses a market or technical need they are usually successful and can be a very exciting ride.

Yes, these are indeed true of many Silicon Valley startups. But, there are many which don't fit this mold. Sometimes luck plays a large part and being at the right time and right place with a good idea is all it takes. It does help to have a successful startup under your belt if you want VC funding. Dr. Petersen's talk illustrated many of the ups and downs of the startup world which are not readily apparent to the rest of the world who read the success stories of Silicon Valley startups like Intel, Apple, Yahoo, eBay, Google, Youtube and others.


No comments:


© 2007, 2008 Madan Venugopal    All rights reserved.