And many hundreds of startup companies have been written about on the pages of EE Times over the last decade and considered for the Silicon 60.
And the list is global. There is no doubt that Silicon Valley is still the largest center of startup activity in the electronics and semiconductor domain, but perhaps not to the degree it was.
One quarter of version 15.1 of the Silicon 60 is headquartered in California, and nearly half is based in the US. After that, in order, come France, China, Israel, the UK, and then more countries represented by one or two companies.
Many of the companies located in the San Francisco Bay Area are there for reasons of sales, marketing, and finance, while the senior executives have roots elsewhere. It was observed that in the last published iteration of the Silicon 60 about half of the US companies had Indians as founders or co-founders although only a couple of companies were actually headquartered in India.
Things are changing. In version 14.0 24 companies were headquartered in California, or 40 percent of the list, and in the latest iteration it is down to 16, or 26 percent, of the list. Does that mean California is getting less entrepreneurial?
Probably not, but it may demonstrate that as the Chinese manufacturing and consumer electronics nexus has become stronger, and as Internet communications become more pervasive, the need to locate a startup in expensive Silicon Valley has diminished.
There have been Silicon Valley-style startup companies in the eastern hemisphere for many years but considerably fewer than in the United States. At the same time it has been hard for many of those companies to achieve "escape velocity" and go on to be globally successful.
But now several eastern hemisphere companies have emerged as global market leaders. There is foundry giant Taiwan Semiconductor Manufacturing Co. Ltd. but also younger fabless chip companies such as Taiwan’s MediaTek and China’s Spreadtrum Communications.
This, together with executives returning home after working in Silicon Valley, encourages the capital investment and entrepreneurship for a new generation of startups in Asia. It helps build the economic success and reinvestment engine that has served California well for so many decades. But such changes do not happen suddenly, as it can take many years to gestate an electronics startup. Of the Silicon 60 v15.1 more than half (33) of the companies were founded in 2009 or earlier, with the rest founded in 2010 or later.
Then there is the argument that venture capital doesn’t invest in semiconductors any more, which also appears to have hit Silicon Valley. It is true that venture capital in the Valley has fallen out of love with the traditional digital semiconductor startup and begun mooning over application software and social media companies.
The reasons are clear. The cost of designing a digital chip at the leading-edge has gone up exponentially, and the ability to create transformational value appears to have diminished. The digital chip company is often expected to write much of the software that will run on the hardware without being paid for it.
Meanwhile, in a pure software business, not only does the company get paid for its wares, but the cost of entry is lower and time-to-money is shorter. But that also explains why the number of would-be players is an order of magnitude higher, thereby reducing the statistical chances of success.
But there is a brighter story for investors prepared to dig deeper into electronics and optical technologies away from the simply faster and smaller. Away from the leading-edge of silicon: In analog and mixed-signal ICs, power and power management, image sensors, MEMS sensors, materials, wireless communications, EDA, and IP cores, the costs are much lower, and the time to money is much shorter.
In the depths of the post-2008 recession we were told that the semiconductor ecosystem was dying due to a lack of startup investment. The industry was becoming frightened because a lack of semiconductor startups was perceived to mean fewer seats for the EDA industry to sell to — and at some point down the road fewer small technology companies for midsized to large companies to acquire at an advantageous price. They might even have to spend more on internal R&D or their own strategic investments in promising startups.
The fact remains that EE Times and other publications have never suffered from a lack of startups to write about. And in 2014 it feels as if we have as healthy a pipeline of startups as we ever did. Impressive entrepreneurs with exciting technology can usually find the money from somewhere or find a way to go around the problem.
The technologies represented in the latest version of the Silicon 60 include: gallium nitride for power and for lighting; chip-scale photovoltaics for energy harvesting; sub- and near-threshold voltage operation of ICs; chip-scale approaches to voltage conversion; impulse radio for location services; 60 GHz communications for WiFi and small cell back-haul; ultrasonic and infrared techniques for gesture recognition; image sensors; MEMS; neural networks; vision and cognitive processing; resistive memory; quantum dot materials; and plastic, printed, and stretchable electronics.
Not all of these technical domains will produce champions or IPOs in the near term, or maybe ever, but we think the 60 companies selected are likely to make the news and be worth following.
From the previous Silicon 60 list it is notable that analog and mixed-signal IP cores developer Cosmic Circuits (Bangalore, India) was acquired by Cadence Design Systems; that Javelin Semiconductor (Austin, Texas) with CMOS power amplifiers for 3G cellular communications was acquired by Avago Technologies; and that Qualcomm has acquired Wilocity Ltd.(Caesarea, Israel), a developer of 60 GHz chipsets based on the IEEE 802.11ad standard, also known as WiGig. That purchase price has not yet been disclosed, but negotiations were reported to be at around the $300 million mark.
Meanwhile a notable casualty from the previous Silicon 60 list isCalxeda Inc. (Austin, Texas), which had developed a server processor based on the ARM architecture. Its demise can be put down to a mix of bad timing and too little money while trying to compete with established players. It came to market with a 32-bit server chip offering, while larger ARM partners have spent more time and more money developing 64-bit server processor solutions.
Those four companies express in microcosm the change discussed earlier: Analog and mixed-signal IP, power, and wireless investments have produced a result for investors while a bleeding-edge digital investment produced a bust.
It’s not a black-or-white issue, and digital companies that can benefit from 28 nm CMOS and finer geometries can be found among the Silicon 60, although often as technology and IP developers and contributors. But it is a factor worth considering as you keep your eye on the Silicon 60.
To view the complete Silicon 60 v15.1, go here.
This article first appeared on EE Times.