Whoever wrote this article must have been away from Silicon Valley and its semiconductor industry for a decade or more.
First, it does not take a fortune nowadays (in the context VC type money) to get a fabless semiconductor start-up off the ground. In fact it costs very little, in the range of $1-2M. It is true that to get a semiconductor start-up to the desired liquidity event can take 6-10 years and can cost tens of millions of dollar of venture money. However, the reason is NOT inherent in semiconductor industry. Most of the time it is due to one (or typically more than one) restarts or resets. These, in turn are typically due to either technical problems with the founders' original idea, or, more often, their lack of understanding the target market and its needs and values, as well as problems with execution. Hence, if VCs did a better due diligence (especially better due diligence on the technical and marketing assumptions of the start-up) they could de-risk their semiconductor investment significantly without resorting to "pre-selling" it as recommended by this author. In my experience most VC-s are negligent in technical due diligence, taking the maxim that success depends mostly on the quality of the founders to the extreme. If VC's do not want to take risk then what is their raison d'etre?
Second, as it is pretty well understood in Silicon Valley liquidity events when a large company buys a start-up company much more often than not shortchanges the engineers whose blood and sweat built the start-up the company. Hence, I am pretty convinced that this model, if it becomes the norm, will slowly but surely will kill semiconductor start-ups, as no engineer in his/her right mind want to work 60-70 hour weeks for years solely to enrich the VC-s.
Well, the differenece in our experience and point of view may be related to the fact that you seem to be a digital engineer and my experience is mostly in the analog/mixed signal area. Design tools are indeed expensive, in the $100k+ area, and you can tape out a chip with a regular maskset and get your 20 wafers for about $120k. So, certainly for $1-2Million a start-up in my area can demonstarte its technology well enough to get a second round funding which enables it to go into production (assuming the technology and the silicon works). Furthemore, I did not argue that all semiconductor start-ups are successful, far from it. If you talk to VC-s they will tell you that only one in ten start-ups is expected to be successful, whether in semiconductors or in any othet field.
A Book For All Reasons Bernard Cole1 Comment Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...