Posts Tagged ‘ EEtimes ’

EETimes.com – Panelists question fabless model viability (Dec/2009)

EETimes.com – Panelists question fabless model viability.

Panelists question fabless model viability
GRENOBLE — Is fabless still fabulous? In a panel session at the IP-ESC 2009 Conference this week in Grenoble, France, panelists discussed the evolution of semiconductor business models and confronted views on whether the fabless model is dead or alive and kicking.The semiconductor business model has evolved, from the IDM model to pure-play foundry, fabless and IP provider, design services business models. Many of these models are undergoing severe issues. “We are talking of a disaggregation of the full model,” said Paul Slaby, president and CEO of Kaben Wireless Silicon Inc., wondering what is coming next.

Paul Slaby, president and CEO of Kaben Wireless Silicon Inc., proposed a break-up of the fabless model. “Time has come to break it down into pieces and get to a semi-fabless model.”

This model consists of a development organization —IP-based design house, outsourced R&D operation, strength in specialized R&D and product development capabilities— and a delivery organization —product-to-market with a sales channel and outsourced product delivery with an infrastructure and a pipeline to market.

According to Slaby, the benefits of a semi-fabless model are the tradeoffs between license-NRE-Royalties.

On the development organization side, he noted that this model avoids raising huge amounts of capital, lowers risks in R&D, market and infrastructure, brings new business schemes such as private labeling, branding and licensing. It also brings scales due to built-in royalties and is investment-worthy.

On the delivery organization side, the semi-fabless model lowers up-front costs, improves financial performances, minimizes risks and expands the product portfolio.

Slaby pointed to alternative approaches from publishing, pharmaceutical and Hollywood movie production models. In the publishing industry, for instance, IC developers are the authors, supply managers/fabless are the publishers and the foundry is the print shop. Another example is the Hollywood movie production that creates, finances and distributes. “Would something like this work in the semi industry?” he questioned.

Today’s reality is that complexity is growing, opportunities to learn are reducing and everything is converged, said Kalar Rajendiran, senior marketing director at eSilicon. “To make a chip, you need 40 expertise and proficiency domains to master. You have 100s suppliers and potential partners to choose. Thus, it is harder than it looks, and the ‘Do it yourself’ is dead. These trends are forcing new business models.”

Moving to eSilicon’s core business, Rajendiran said “eSilicon’s Value Chain Producer (VCP) model has it all today so we can address the industry and help it grow.”

Introduced at the 45th annual Design Automation Conference (DAC) in Anaheim, California, eSilicon’s VCP model consists in providing a comprehensive suite of design, productization and manufacturing services, enabling a flexible, low-cost, lower-risk path to volume chip production.

To the question “Is the fabless IC model alive and vital”, Stan Swirhun, senior vice president & general manager, Optical Products Group, Zarlink Semiconductor, answered “yes and no” as it depends on size, business focus and maturity.

There is a decreasing number of fabless startups, and Swirhun said VCs are increasingly taking a low-capital vision and, eventually, funding for fabless startups continues to decline.

To succeed, small and mid-size fabless companies must reduce operating risk by owning a great technology or market, by focusing on a long-lived product and by focusing on a core capability. He also encouraged to serve existing customer relationships, to rely on narrower internal capability and hired experts, to rely on partnerships for differentiation and cost benefit, and finally to continue to explore ‘own less/risk less’ business.

Talking in the name of VCs, Jean-Philippe Gendre, investment director at Emertec Venture (Paris, France), explained why it is so challenging to invest in early stage companies. “Life cycle funding might require raising $50 million, and exits above $200 million have become very unusual. This means that expected returns for VCs is limited,” Gendre said.

He continued: “Many things can go wrong, and every error is costly and time consuming. In addition, sales cycles are usually long, and you may burn $50 million before you get a product to the market.”

Gendre then highlighting some prerequisites for success, and his first tip was to “have an excellent team in terms of execution.”

He further explained: “It is important to have a scalable solution so that you can derive several products and draw from this investment. So, it stretches the lifetime of the investment.”

Second on the list was strong VC syndicate. “It is key to enlarge the VC syndicate to refinance the company if it makes sense at some point,” Gendre noted.

And, third, it is essential to invest on very dynamic and sizable markets as he said “we need to have growth perspectives.”

To lower VC risks, Gendre highlighted the need draw the ecosystem. This goes through some support from vendors —EDA and foundries— so as to adapt pricing to early stage companies and from customers to facilitate the first design wins and share risks. Another way is to find other funding options such as subsidies.

Ten emerging technologies to watch in 2010

Peter Clarke
(11/18/2009 11:40 $ EST)
URL: http://www.eetimes.com/showArticle.jhtml?articleID=221900117 LONDON — EE Times has compiled a list of emerging technologies that we think will be worth watching out for in 2010.
Recessions are the times of change when R&D investments get pushed to the fore. It is well known that when markets and prosperity return they never return in exactly the same form that they went away.
We have deliberately favored the hardware- and physically based side of the technology landscape, although software is also likely to increase its impact and importance in 2010.
There are also some technology trends that are so self-evident and long-term that we have not listed them. We would include amongst these the need to reduce power consumption and the need to pursue low-carbon and reduced materials content solutions. We see these as drivers for some of the more detailed technologies we list below. We don’t claim to have a perfectly accurate piezoelectric crystal ball, but some technologies and some technology providers are going to change the landscape in 2010. The ten technologies listed below, in no particular order, might just be part of our changing times.
Continue reading