Twinkling of Semiconductor Industry in the Gloomy Economy

I attended the CASPA/GSA joint conference yesterday. The purpose of the conference was to "rejuvenate our semiconductor professional community" and to "take a breath and refresh our mind." The main speakers were from GSA, Marvell, Cadence, GSMC, and Deloitte Consulting LLP.


GSA showed revenue predictions from 11 industry research analysts. Most analysts expect at least a 2009 revenue decline of 20% over 2008. There were other slides showing companies with positive projections or strong cash positions to help them outlast the downturn. "Cash is king."

Jerry Sanders and T.J. Rodgers once said, "Real men have fabs." GSMC altered the quote slightly by stating, "Real (stubborn) men (still) have fabs. Others have profits." For those few companies that have the business model to support owning a fab, there is an advantage of improved yields and access to technology not available to others. For the majority of companies owning a fab is difficult to justify. I've felt for a long time that owning our own fab has no advantage and burdens us with high costs. Our fab has suffered with dismal utilization and uses an older process. Foundries such as TSMC and UMC provide fabless companies with access to the latest technology nodes. By eliminating fabs as a competitive advantage, fabless companies need to find other competitive advantages to be successful.

China is a place of opportunity for many people. They have many engineers but lack discipline and methodology. There are chips taping out without DRC checks and sometimes people even forget to integrate IP. With stories such as this I believe that those with the relevant experience, desire to lead, and ability to move to China should do well for themselves.


There were some heated discussions during the conference especially when personal career paths were shared. Someone in the audience took offense when a speaker said that there were too many local engineers and that he enjoyed his job a lot more after moving to sales after 17 years of being an engineer. The audience member disagreed that there were too many engineers and blamed outsourcing. He stated that Silicon Valley was the source of innovation and should remain that way.

Other speakers mentioned that on average people change careers 3 times in their lifetimes and people needed to find their strengths and be open-minded to opportunities. They agreed that careers are sometimes beyond an individual's control and in part are dependent on luck. One speaker first worked at a startup which was bought by Synopsys. After the proven track record he was able to raise money from the same VC and started another company that was sold to Cadence right before the technology bubble burst. He was able to have a career as an entrepreneur in startups, as a partner at a VC firm, and finally as a Sr. VP, Chief Strategic Officer, and acting CTO at a big company.

Another speaker earlier in his career had an opportunity to join either Intel or Fairchild Semiconductor. Intel was still a small company at the time and he might have been able to retire early with that choice. However if he had picked Fairchild Semiconductor, which later became a Superfund site, he might have been involved with lawsuits because of groundwater contamination from leaking storage tanks.


One speaker spoke of the need to better control operations to avoid the hockey stick effect. Many companies try to book sales quickly in the final weeks of each quarter in order to meet wall street expectations. They do this by giving up something such as accelerating cost reduction schedules to convince their customers to take additional inventory. He noted that both parties lose with this transaction. For the vendor, they just accelerated the reduction of ASP and their profits. For the customer, they need to handle additional inventory even though it might not be used right away.

In addition, the speaker stated that the vendor has to manage inventory and predict when orders might come in. There are long lead times for the semiconductor industry and a limit on rush production capacity. Often a vendor must issue risk orders to be able to fulfill expected orders. By pushing customers to buy inventory that's not needed, there is lower visibility of demand. The vendor created artificial demand at the end of each quarter and can't tell when real demand might appear. Ideally factories should be run smoothly and production levels should be stable. He gave an example of diaper factories that also don't run at the same levels throughout the year. Babies do not suddenly consume more diapers at the end of each quarter yet diaper production might increase during that time. Inefficiency and lack of visibility in the supply chain is costly. However I think the end consumer might benefit. For example customers at costco might soak up excess diaper inventory by taking advantage of coupons and lower prices.

Final thoughts

Speakers agreed that innovation was the key to improve the industry. Possible areas for investment include 'green' technology, analog, home automation, communication, automotive, and healthcare.

One speaker mentioned that many people change careers either when the economy is very good and new opportunities present themselves or when the economy is very poor and there are layoffs. I wonder what my future holds... still looking for my twinkling star...