I received a plethora of e-mails regarding my recent article, entitled "Opinion: Semi IP sector is a lost cause"
Then, I asked for the IP vendors themselves to respond to a set of questions. Here's one letter to the editor:
To the editor,
"Having been there and done that", I am following this lively debate on semiconductor IP business with great interest. About 10 years ago I started up and ran a small design company developing and selling what we would call "commodity" SIP products such as 8-bit processors, I2C, USB interfaces, etc. Later on I was the CEO of ATMOS Corporation, a venture capital funded embedded memory company, which was in "star" IP business.
I have experienced both the low-end and high-end of the SIP business and I can attest to most of the comments made in the current debate. Without repeating them I would like to bring a bit of an entrepreneurial angle to the discussion by way of the following points and observations.
1. The state of the semiconductor IP business seems to be reminiscent of the state of the automobile industry in the beginning of last century. In 1893 Duryea brothers launched the first 1-cylinder auto in the US. At the time the horse & buggy dominated transportation. In the early 1900s there were 100s of auto manufacturers building custom-made cars which were a luxurious novelty -- one model even offered electric curlers in the back seat for on-the-go primping. Cars were unreliable and expensive (about $1,500, 2x annual family income) and the public resented them.
The industry was small and unattractive. Then in 1908 came the famous Ford's Model T built "for the great multitude, constructed of the best materials" which revolutionized industry. It came in one model and one color (black) only but it was reliable, durable, and easy to fix. Furthermore, it was priced so most people could afford it ($609-$290 versus horse-driven carriage (competition!) at $400).
Ford's success was underpinned by a profitable business model -- hear, hear, semiconductor folks! The cars were highly standardized, had limited options, interchangeable parts. To achieve that Mr. Ford innovated with a revolutionary assembly line which allowed him to replace skilled craftsmen with ordinary unskilled laborers who worked on small tasks faster and more efficiently. As a result he cut production time from 21 days to 4 days and labor hours by 60%. With lower costs, Ford could charge price accessible to mass market.
As a result sales exploded and by 1923 majority of US households owned an automobile. The rest is history.
Now, it is easy to be presumptuous that all we have to do is follow his approach but the parallels are obvious.
2. The SIP industry -- pursuing the "chipless model" -- is quite diverse consisting of literally of hundreds of players, most of them rather small, functioning as cottage-industry, and just a handful large enough to attract the interest of the financial community. This chaotic diversity, again reminiscent of automobile industry in 1900s, contributes to the vastly polarized views on the viability of the industry as expressed in the current debate. Certainly the largest players managed to develop sustainable business models which they pursue with reasonable success as reflected in their market caps and valuations.
3. From the entrepreneurial angle, semiconductor industry these days is a tough place to play. It is a maturing industry with enormous capital requirements to get anything sizable going. Even the fabless model these days requires typically $50-100M in VC money to get anywhere near sustainable operation. Then, of course, there is a considerable risk of the total wipeout due to technology problems or market timing as evidenced by tens of examples from most recent days.
Indeed, from the business or investment perspective, you would need to be a really brave soul or truly desperate to start a fabless semiconductor company from scratch these days. There's got to be an easier way to make a living! And indeed, this kind of thinking leads many of us with entrepreneurial itchings to found startups pursuing various mixes of IP business and design services.
In a classic, Darwinian way they claw their way up like those new-born turtles crawling to sea on the beaches of Galapagos. Many of them will die before they reach safety of profitable business models but the best of them will become successful, no doubt with far less hard-to-get venture capital. Their success helps to shape and re-shape the microelectronics landscape.
4. Just like the automobile industry, the semiconductor business is changing, maturing and consolidating. Who knows, we may even end up with just three large semiconductor firms in North America! But it does not mean that there is no room for new innovative semiconductor businesses to pursue. We may not be able to get into chips-cranking giant foundry business, just like it is impossible today to start from nothing a new car making plant, but there is a booming parts making business, with multi-billion dollar players such as Magma International, supplying those plants.
The SIP industry needs to consolidate to create a few anchor-type players that will make the whole IP ecosystem viable. The consolidation process creates a lot of value -- see the recent $147M ChipIdea acquisition by MIPS. So, let's roll up the sleeves and play!
C. Paul Slaby
Kaben Wireless Silicon Inc.