@eewiz- I like the idea of EDA vendors and foundries throwing skin in the game by providing reduced cost tools and wafers to startups with the agreement that they will pay higher prices when they reach a certain threshold. It's important that they ensure a steady stream of new customers, and nurturing promising startups is a way to do that. Perhaps some of the smaller EDA vendors and foundries will get the ball rolling on this as a way to get their foot in the door at what could be big customers down the road.
@vaapb, many good points...I belive EDA companies are trying to address the issues you brought up, I know from personal experience a start-up company that was paying little, if anything, in their first year of IC design...the start-up eventually went belly up so the EDA company in quetsion likely lost the money on that "sale"...
There is a difference between web and silicon start-ups, number of them! We can't really expect millions of silicon start-ups wordwide, that would not be good either...
In the context of this GSA initiative a useful suggestion might be to use some funds to lease EDA tools for start-up purposes...Kris
EDA tooling costs are prohibitively expensive and present a barrier to entry that just doesn't make sense. It is a barrier that this industry can do without - if it has a prayer of reviving here in Silicon Valley.
It always saddens me to see ivy league EE/EC grads get into web programming - after all that hard work in earning their degree; what a waste. I suspect that is the case because access to good EDA tools is simply beyond their budget for experimenting in a serious way.
One reason the web start-up companies really took off is because of the hosted model and plenty of inexpensive/free tools.
A better idea would be to invest all that money into building a cloud with the best of EDA tools and processes and make it affordable and accessible.
This will allow designers/entrepreneurs with great ideas and passion focus on what they do best without having to worry about EDA companies eating their lunch money even before they get started.
I understand that EDA companies also have to make money to survive. But their current pricing models I believe is doing more to kill their own ecosystem instead of growing it.
If I had a billion dollars I would just buy the software from these companies and opensource their software - like what Google did with ON2 codec.
Markets and opportunities appropriate for startups or even small cap companies have changed - driven by the onward development technology.
There are still semiconductor related opportunities, but great company ideas are not as common as people would like to believe. It requires a very compelling product offering, accessible market, and strong team to succeed at any startup. I've personally started and profitably sold multiple companies and currently advise and serve on the board of others including a semiconductor company.
Apparently the "obvious" semiconductor opportunities are very capital intensive, but in my opinion the real opportunity for startups are in special purpose devices (mixed signal, RF, exotic uses) and use of enabling technologies to implement higher level value to the market.
Few, if any, company that raises and spends $100M while private is going to generate a reasonable return to the investors. It just does not make sense to do so when troubled, but revenue generating companies can be acquired for far less and repurposed for turnaround and growth.
In general to design a moderate complexity FPGA and put it on a board for prototyping you will need a half dozen or more very talented engineers for a couple of years. Of course that time and man power estimate can scale quickly if you are getting into custom IP. Tools, test equipment and space add to the expense.
To me the bigger problem is once you get an FPGA prototype then you need to convert to ASIC which takes 6 to 9 months conservatively.
Now after ~2.5 years you have your new wonder chip that you need to get OEMs to design in. The product life cycle from design-in to production for consumer electronics can run from 18 months to 36 months or more.
Now your new OEM partner with your chip in their product has to go sell it to their customers. They may have field trials, further certification, documentation and support teams. And most importantly market acceptance.
So from idea to end product you are talking 4+ years and that is if everything goes well. This is the funding problem.
David: One idea that many start-up try to follow is to design some IP blocks not the whole chip. That gos toward smaller complexity and number of people required. The problem is that you still have validate that IP thru silicon fabrication...Kris
Rise in costs for a semi start-up is dominated by the combination of increased product complexity and the number of people needed for development.
Only changing the biz model to reflect this change will decrease these two factors and reduce the funding needed to go from idea to profit.
Good Intention, bad solution. 100m$ angel fund is like is cup of water for a man dying in a desert. it will finish of fast as Frank said. what should be done is lower the costs by negotiating with fabs and eda companies to support a startup model.
ie. SYN/CDN/MENTOR should give access to semi startups at very affordable prices. For giving this they can add contractual terms like buying the EDA company software at full price when the company becomes succesful. Or they can take a small stake in the startup and give free tools, sort of a mini VC role by the EDA companies. Or atleast they should software-as-service model for startups.
Same thing with the foundries. Negotiate with the foundries for lower costs for startups, on condition that the startups will continue with the same foundry when they becomes successful. Again, startups can trade a small stake in the company to foundries in-return for low prices.
All these things are theoretically possible even now, but having an official forum and a pre-negotiated agreements with a range of prices/equity will help the startups.
A suggestion to startups is to focus on niche markets with low competition, where older process also may be enough to demonstrate ver 1,2 of the products. "Primesense" who made the 3d camera chips for XBOX kinect is a good eg.
Interesting - but I think this will not work out. None of the angels will intentionally put themselves in this brutal and unforgiving sector. Our field has become stymied because our innovation is no more fluidic. It will take a lot of new ideas to compete with the clones Chinese are making today. It is a tough business people because China has found a way to devalue anything you make.
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 ...