The 6 Must-Do’s for Semiconductor Fabless Start-ups

Getting funding for a semiconductor startup is not getting easier. According to CBInsights, there have been fewer than 20 Angel/Seed and Series A deals in the US in each year since 2009.


So where is all the funding channeled to? In 2013, close to a quarter of all VC funding was directed towards internet companies. (Preqin, June 2013).  If you were a VC, you’d probably do the same. Internet companies are more agile. They can easily change their product, business model, and transform the entire company direction overnight. On top of that, the investment is a fraction of what’s needed for a semiconductor startup and let’s face it –  Internet companies are a bit sexier.


With  maskset prices only going up,  raising money is not going to become any easier.  So what are you options as a fabless startup? While we can’t solve the funding challenge, here are a few tips that may help you limit your expenditures.


1. Collaborate with Foundries and Assembly Houses

As a new start up you must be a in a constant look for new competitive edge whether it’s price or performance,  power consumption, there is a lot to gain in the production phase as well. New technologies in packaging can reduce the total device cost and better collaboration with the foundry can give you better understanding on how to optimize your design for better performance.


On the other hand, assembly house and foundries are looking for new requirements in order to drive new products and technologies and without having direct contact to the market they are missing real-life demands to enable new features.

log1 log5 log2

2. Partner with IDMs

While qualcomm, broadcom and other large IDMs are looking for getting more market share and beat competition – perhaps you can help them. By using IDMs distribution channels you can get closer to your customers and even more close to get some cash.


The tricky part is to find the IDM that needs your technology and needs it now. Try to connect to business development folk to see if your offering make sense for their sales strategy. There have been several start-up and IDMs engagement in the past that ended up in a happy marriage.


3. Find Great Semiconductor Suppliers

To lower the risk and reduce time to market, you need great partners. Use AnySilicon search tool to identify vendors that can support you quickly. Some suppliers offer very professional solutions and services together with creative business models that can help with reducing the upfront investment. A combination of technical and commercial engagement could be a key to your success and really value add for a small startup.


AnySilicon’s directory is a good place to start when you are kicking off a new project. The number of vendors in the database is increasing daily and offer huge diversity, from ASIC design, wafer supply, package design to FIB services. Everything you would need to realize an ASIC.


4. Engage with Customers before you Start the Design

We have been involved in many semiconductor products for many years and seen big mistakes consistently repeated. One of the biggest mistakes is about product definition. Engineers and managers believe they know what customers want even without meeting them or talking to them.


In any company, and especially in hardware-based companies, where there is a lot of money involved with product re-definition – changing a product means money. For fabless companies adding a feature which is hardware-related (vs. software related) can cost a lot of money. Adding functionality into the silicon will consist of: design and verification of the ASIC, a new maskset, production test updates and sometimes changing the package type. This requires a huge capital investment and at least 6-12 months of delay


5. Think Vertical

There are many examples of fabless companies changing their business models and ending up selling a product instead of a chip. The motivation behind this step was mostly because they could not find any customers for their silicon — so they decided to design, build and sell the end-product themselves.

There are great examples of chip companies selling end products. SanDisk is one of them. We all know them as the USB-stick company, but actually the core competence of the company is ASICs. The end product is different and obviously generate much more profit.


This means that Instead of selling chips, you may want to think vertical – can you sell an entire product based on your chip or perhaps a module?


6. Buy IPs, Don’t develop them


Don’t, please don’t develop IP blocks, if they exist, buy them.

The three top places to go today are:

AnySilicon.com     (we have IP listing since Jan-15)




Recent Stories