The rise of China's fabless industry

6/15/2011 8:10 PM EDT

What if only Samsung has the political freedom to serve Chinese fabless; tsmc and gf would obviously come under fire.

6/15/2011 9:36 PM EDT

All true.
Taken to the next level, it raises questions about the future of the U.S.-based fabless semiconductor industry, where "revolutionary" technologies and products will come from, and the future for U.S. employment in the fabless semiconductor industry.
First - few U.S. VC's will invest in a fabless semiconductor company today beyond an "A" round. Current levels of investment limit a potential fabless start up to the development of a proof of concept and a quick flip of the company. No money is spent on developing the infrastructure (Sales, Marketing, Engineering, Operations, Marcom, Support, Sales Channel development etc.). No infrastructure, no U.S. jobs. Job growth is where the fabless company is and the final product is manufactured - China.
Second - The large, mainstream Semiconductor companies are now more focused on specific market segments or niches and are very protective of "their" market space. They leverage existing customer relationships and offer evolutionary products enhanced with new technology IP (see "flip" above), and lower prices. This make it difficult for a new company to insert a new product idea into "their" market space.
So, Is the fabless model now a commodity Chinese-based industry with evolutionary extensions of current products and a focus on simple, higher levels of integration?
What will fuel the investment and development of revolutionary technologies?
Is there a future for the U.S. Fabless semiconductor industry?

6/15/2011 11:31 PM EDT

Good points idea guy. It does look like the US will no longer be the only place where new ideas happen. This will take us another notch down in the food chain.

6/16/2011 12:37 AM EDT

Meanwhile tsmc and gf eagerly look forward to US fabless business...

6/16/2011 2:18 AM EDT

Mirror, Mirror, tell me who is the cheapest of them all? I think no fab will be able to beat China in terms of cost AND efficiency. Sustaining fabs in US is too costly. Many companies including Texas Instruments, AMD have gone 'fablite' if not fabless. The list of companies in GSA, erstwhile FSA (fabless semiconductor association) is only going to get consolidated through mergers and take-overs. I dont know how companies can/will go beyond 20 nm! There are some ideas about 3-D silicon, courtesy Intel but I guess it is still in 'academic' phase.

6/16/2011 11:52 AM EDT

It is amazing. I do not remember reading about the first round of investment in Chinese fabless companies, "...not working out...". I guess it always makes better print to talk about how the sky is falling for the U.S. and China will take over the world (hey, it got me to read this article). Yes there will be fabless in China. Will it disappear from the U.S.? No, it will adapt to the applications it does best (and China will make what it makes best). If we become like Britain, it will only be because we have adopted the same socialist government-run economic policies that raise the costs of doing business here. China's revolutionary solution was to create zones temporarily free of most government interference.

6/16/2011 11:11 PM EDT

I think the story of 'US' being the hub of innovation is running longer than any soap. It might have been true in early days when it took some time for the 'knowledge' to cross the oceans and there was an upfront 'lead time'. Now in the 'internet era', how can anyone say that 'innovations' can happen ONLY in US and not in Europe or Asia? It actually depends very much on whom the innovation depends on. It is difficult, to put it mildly, for US companies to innovate on what the Chinese might queue up in the night to buy, forget just like or feel good-to-have.

6/17/2011 8:41 AM EDT

Within any economic analysis of product development in the USA is the design resource expense which is substantial. Designers ultimately take years to develop their skill set but with time China teams will have the same. So the way one cultivates that is create more design efforts. I have been in many executive meetings where the topic is how do we end the monopoly of design expense in the USA, well that is the next round of outsourcing. I would expect the crystal towner of design to topple in the next 5 to 10 years in the USA, and become flat as the world equalizes expenses for design value across the globe.

6/17/2011 9:04 AM EDT

@BLinder

You may be right about the established crystal/ivory tower(s) of design expertise collapsing in next five to ten years.

But why would that result in a flat landscape?

The equipment making all goes on in China. Much of the chipmaking is done in Taiwan, Korea, Japan and more will be done in China. Surely the expert knowledge will continue to migrate to, and develop in, China and southeast Asia?

Or is there another prognosis?

6/17/2011 7:29 PM EDT

@ Peter, What I meant by flat was global wage equalization. Here is the problem, as the mask cost goes up (500k +++) for given technology nodes, some bean counter is going to ask how do we cut cost. The mask makers are not going to budge, so that leaves the design resource salaries. Please note, I have worked with design groups that spin 3-4 full mask sets prior to final release. Really hurts the pocket book!

6/18/2011 2:40 AM EDT

If there is a so-called "design crystal tower", it is not unique to "design" or to "engineering" or to "technology". I perceive the "crystal tower" is referring to self-interest. The self interests have led to an overall high cost for technology or any kind of development, relative to other places.

For example, if the expectation is to have a large fancy home, an engineer would have to orient his/her priority accordingly. If the expectation is to have leisure on priority, an engineer would be expected to put leisure on priority vs engineering efforts. If the expectation is to consume, an engineer would be as wasteful as any other person.

An engineer's or a manufacturer's work is easier to be out-sourced to another place than the work of a lawyer,or a doctor, or a politician, or a homw builder, or a taxi driver, or a hospital worker, or a policeperson, or a teacher, or a - -.

The basic issue is not one of technical or innnovation ability but rather priorities of a society. A priority change from one of consumption to one of infra-structural building , from reward for self interest protection to for social contribution. Measure infra-structure/industrial expenditure index instead of consumption index. Provide incentive for saving/conservation as opposed to expenditure/environmental-impact.

6/18/2011 4:32 AM EDT

@StephenWL

What you say strikes a chord.

It is notable that in the West the numbers of young people who want to study science and engineering is reducing, which perhaps reflects a lack of interest in infrastructure building and development of the country.

Meanwhile in southeast Asia and elsewhere where there is, I perceive, still a lot of interest in infrastructural building STEM (science, technology, engineering and maths) are popular.

Another way of looking at it is the difference between the questions "What can the country do for me?" and "What can I do for my country?"

6/19/2011 4:46 PM EDT

It's inevitable. We might question the pace of it (I have to say that even I, a self-confessed free marketeer is surprised with Western Government's complete resignation to this trend) but in the end the inevitable will happen. The Western answer should be constant innovation. There is no other way, I am afraid. Capital will always have the last word.