How to start a semiconductor company: the story of GreenPeak

Launching a hardware innovation such as a real electronics product is said to be harder than launching a software product or service. But how much time, effort and money is really needed? We asked these questions to a veteran in the wireless communications world: Cees Links is CEO of GreenPeak, a ‘fabless’ semiconductor company headquartered in Utrecht. Cees was involved in the invention of Wi-Fi, and with Green Peak he is now making chips for low power ‘ZigBee’ communication.

How will ZigBee change our lives?
ZigBee is a low power communication standard, and is already being used, for instance in TV set-top boxes, remote controls, and door/window or other building sensors. With ZigBee you can make the battery last as long as the economic lifetime of the device, so you never have to change batteries. This is a cost advantage and an environmental advantage, and if we can bring down power consumption further the battery size can be reduced or more power can be made available for more sensors or bigger displays on devices.

Zigbee demonstrated in and around the house. All kinds of new smart devices will need to communicate wirelessly with each other

In the beginning we were aiming to completely remove the battery, and make devices work on some form of green energy like solar. We discovered however that a small battery or power storage is needed to make the device work anytime. In the end it made more sense to charge to power storage environmentally friendly and efficient in the factory, and save resources by leaving out the solar panel altogether.

Like other midsize electronics companies, Green Peak is ‘fab-less’. What does this mean for the business model?
We have an international supply chain. We design the chips here in The Netherlands and in our other office in Belgium. We buy wafers in Taiwan and pay a Taiwanese factory to put the chip design on it. A Chinese factory then separates the chips and finishes the product. We store and distribute them from our distribution center in Hong Kong. So we do sell and deliver our own products (the chips), it is just that we rent the factory on an as-needed basis.

What are your key markets?
The early adopters for Wi-Fi were in US and Japan, and the same is true for ZigBee. The US is a huge market so it is important to get a standard approved there first, and the Japanese are just very enthusiast about getting the latest technology. However ZigBee is only this year coming to the Netherlands: UPC has recently launched their new Horizon set-top box and it uses ZigBee.

Have you considered other business models?
Not really. It does not make sense for a small or midsize company to own a chip making plant. In theory you could, as an inventor, have others produce your inventions and only ask for royalties, but I have not considered it for GreenPeak. We have the vision and drive to make this a successful product and decided to make this happen.

How much time does it take to build a semiconductor company like Green Peak?
In 2011 we had for the first time more than € 1 million in revenue. So we now have significant sales and can fund much of our growth and running cost from the revenue. I would say that it will always take 6-7 years to reach commercial success for RF-semiconductor company. With Wi-Fi, we started in 1988. In 1991 we had a first prototype, and in 1996 the standard was published. Wi-Fi however only took off in 1998 after Apple’s decision to use the standard. Perhaps there are some niches were innovations can be launched faster, but most chip companies follow this 5-6 year roadmap.

How much funding was required by Green Peak?
In total the investment was about € 30 mln. We started the company on one of the last days of 2004. We could have saved ourselves € 150 Chamber of Commerce contribution by waiting a few days, and in those days this was a significant sum for us, but we only realized this after the fact.

In the year 2005 we obtained € 1 mln of seed funding and this was sufficient to get us through the first year. In 2006 we received the first tranche of our Series A funding of € 4mln and in 2007 we received the second tranche of this funding round of € 2mln.

What did you do and achieve with the first € 1 million of funding?
After one year and one million, we had built a team and a vision. You cannot expect more than that, since it takes multiple years to build a chip product. Investors invest in the overall vision, and a team with experience to reach the vision.

Did government subsidies help in building this company?
Yes. We received the first ‘Uitdagerskrediet’ from the Dutch government in 2006, and in 2010 we received a new subsidy “Innovatiekrediet” to develop the next product. The subsidies alone are only a small portion to cover all development costs, but they really help in closing the finance gaps.

Has it become easier in the last ten years to launch a product?
Not for the kind of products that we build. You need time and effort to design a chip prototype, get it tested and approved, produced and make it a standard. Since our chips do communication, they are a mix of analog and digital circuits. For the digital parts we do use programmable chips (so-called FPGA’s) and this makes it slightly easier to build a prototype. The difference for a purely digital chip would be 1.5 year instead of 2 years, so you still need 5-6 years from start to € 1 million revenue.

Do you think crowdfunding initiatives such as kickstarter will make it easier to launch innovations?
For software companies definitely; for hardware companies I am not immediately saying no, but I do not see how it would work for us as a semiconductor company. The problem is that the development cycle of 5 years is too long. If you campaign on Kickstarter you need to offer your backers a product much faster. It is a good platform for other real products, especially when combined with rapid manufacturing techniques such as 3d-printing. However the semiconductor market we are in still is and is likely to remain a market with high barriers of entry.

Supposing two student entrepreneurs have a brilliant electronics idea and want to start a company. Would you recommend them to start a fabless semiconductor company?
That is a question that I cannot really answer for them. There are high barriers of entry, compared to the software market. This is not necessarily a bad thing: the high barriers will also keep out competition once you are in and the potential profit is large.

What they should realize is that you need more than a good idea to start. I would recommend them to get an experienced team like we did, or start working first to get experience. To get these people on board, you might have to give away most of the shares, so your aim should really be to have 10% of a large company rather than 90% of a small one. The value of the idea is often overestimated. Execution is also very important.

What else do you need?
You need perseverance to succeed, and resourcefulness. You need to keep focus on you grand vision (the 50 billion devices with your chip) while at the same time being creative in finding the best point of entry for the market . The problem is finding the first application: how are you going to sell the first million chips? You need to find early successes.

If you have this experience, perseverance and resourcefulness, there is no reason not to go and try to start a semiconductor company.

How important are patents in your business?
The patent process is very important for us. It is important to see it as a process or funnel. The total cost from first application to getting a patent in several market takes a few years and costs € 100.000 in total. The patents only become valuable if they have been hardened in court: if they have been challenged and found valid. Patents are thus expensive for a startup company. We have applied for about 25 patents with Green Peak, and for about 10 patents we have completed the patent process. Sometimes we only apply for a US patent, in other cases we also or only do Europe or Asia. You really need a strategy here to minimize costs while maximizing the value or protection.

What are the next innovations for Green Peak?
We have many ideas, but have found out that it is also important to have focus. We explore all the ideas we have in our own focus area, that of low power wireless communications. In this area we can add a lot of value, and there are many applications yet unexplored. One area we are not exploring yet are electronic price tags for super markets, to have shelve items report their temperature and condition.

This entry was posted in CEO/founder interviews and tagged , , , , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s