The Future of Innovation: Now, It Gets Complicated, says ITRS Panel
An elite group of executives from across the semiconductor value chain convened at SEMICON West 2009 for the ITRS Executive Keynote Panel, “The Future of Innovation: Implications of Today’s Trends on Tomorrow’s Technology.” Perspectives from the IDM, equipment, materials, packaging and end customer segments were represented. Each leader spoke of today’s challenges, emerging trends, and the implications for the future of innovation.
The global economic crisis, which has created some major potholes in the roadmap process, was central to each of the executive talks. Panel members, moderated by SIA president George Scalise, took turns addressing the historic confluence of the present economy, changing markets, integration of the supply chain, and green opportunities, and provided insight into ways out of the downturn.
Moderator George Scalise, president of SIA, spoke first on behalf of the entire industry, noting that the high level of technology innovation required to compete, coupled with the lack of growth of revenues due to pricing pressures, indicate that we may be approaching a “tipping point.”
A look at industry profitability substantiates his point. Semiconductor operating margins have been in decline since 2005. Although OEMs have successfully harnessed their buying power and extracted a great deal of value out of semiconductors, it hasn’t translated back into company earnings. And, naturally, the economic downturn makes this worse.
In Scalise’s view, we are already seeing the impact. Semiconductor makers have sharply cut their Capex, which is projected to be about a third of what it was just two years ago. The semiconductor industry revenues are expected to fall about 20% in 2009 versus 2008. Other elements of the supply chain will be affected just as badly, and in some cases, even more so.
Due to prolonged negative gross margins, memory companies’ dollars and percent of revenue spent on R&D has finally started to fall. Former IDMs have cut out most of their R&D programs and foundries have cut back on their spending, too. A number of companies have either been bought or gone out of the business.
“Make no mistake, we are approaching a tipping point where traditional, go-it-alone strategies won’t work”, said Scalise. “We must find new ways to collaborate.”
The Necessity of Partnership
The second speaker, Mike Campbell, SVP of Engineering at QUALCOMM CDMA Technologies, spoke of the economic woes facing the cell phone industry. That industry will contract in 2009, down 300 to 500 million units from one year ago, with a TAM of 1.2B phone sales.
While the middle tier is shrinking, low-end phones sales for emerging markets like China, India, and the Philippines are growing dramatically. High-end wireless is growing, too, with increasing numbers of applications. This requires technology innovations to ensure low cost, low power, small form factor, and high performance.
In Campbell’s opinion, the Integrated Fabless Manufacturing segment needs to continue collaboration across the value chain, co-creating the architecture, design, silicon, and packaging. Integration of new elements— such as MEMS, 3D, high k, low k, TSS, and embedded die-in-substrate— require R&D to be pushed back to include early path finding and pre-manufacturing.
“Far more partnerships exist now between the fabless marketplace and foundries” as a response, said Campbell.
Devastation and Hope
Mark Durcan, Micron Technology, presents at SEMICON West at the ITRS Roadmap meeting.
“The downturn has been devastating,” opened Mark Durcan, president and COO of Micron Technology. Year-over-year growth in Capex for the semiconductor industry dropped from heights of more than +60% five to six years ago, to below -60% in 2009. Poor end demand for the customers and difficult times in the customers businesses are to blame.
According to Durcan, significant damage is being done to the supplier base for the first time in a long, long time. Smaller suppliers are especially impacted, as they are forced to spend proportionately more on R&D.
Because of its acceleration of Moore’s Law, NAND will run into problems first, facing challenges in process and cost. Moving to 20 nm over next two to three years will be impacted by difficulties in ability to continue to scale conventional lithography methods without adding significant complexity. “It doesn’t mean we stop,” said Durcan, “It just means we hit new limits that drive new R&D constraints.”
New processes are generating smaller, more gradual revenue peaks, slowing the migration to new technology. And considering that consolidation of the equipment suppliers is exacerbated by the current situation in the marketplace, a new model is needed for how suppliers, equipment, IDM and foundries interact.
When faced with increasing development costs for new technologies and a more difficult technology roadmap, Durcan feels it’s an inevitable conclusion: The industry must drive to codependence and accelerate the rate in which partnerships and strategic relationships are formed.
Finally, Durcan decried the “insidious” and damaging role of ongoing government intervention on the marketplace, seeing it as a “significant negative force” in the business.
Unprecedented Challenge and Trust
Douglas Neugold, president and CEO of ATMI, explained the issues from the materials side of the equation. Most devices now use more than 60 materials. Logic process steps have increased from fewer than 700 to over 1,000 in the move from 130 nm to 32 nm. In the pursuit of relative gain in performance, conventional scaling is still critical, but the materials those devices are made out of are even more important. Neugold believes that integration around those materials represents an unprecedented challenge.
Feature size defines profitability. “Nobody can hide from this,” said Neugold. “Most people make money with the most advanced devices.” At the leading edge, this challenge must be accepted.
These days, Neugold lives in an application-specific materials world. They have developed 65 solutions, each just a fraction of the market they once were. He said that leverage is virtually impossible.
To address this, ATMI is working closely with Intermolecular to develop high-productivity combinatorial science-based research for materials to apply in the semiconductor industry. In doing so, their pace of learning is “revolutionary.”
ATMI realized they needed to rethink the business model and change their historic approach, like sharing with partners a lot earlier in the process. Neugold asserted, “If we’re going to talk about change, trust is an enormous part of that driver.”
Compelling Data for Change
As Thomas St. Dennis, senior VP and GM for the Silicon Systems Group at Applied Materials, sees it, a “real crisis in the equipment industry” exists. The impact of current conditions, in addition to the some of the longer term projections, presents a “compelling set of data for change,” he said.
There’s a need to increase R&D at this time to cover the evolving and diverging parts of the processes. Yet, the total R&D spending by the top 10 equipment companies has decreased by more than $750M from 2007.
“We need to invest more,” says St. Dennis. Yet, at the same time, “we’re already investing more than we can afford.”
The industry needs new thinking and new business models throughout the supply chain. He stresses the need for collaboration across the industry and between companies within the industry. “In the last year, Applied has done things that two or three years ago, we would have not considered,” St. Dennis revealed.
No Middle Ground
Finally, Tien Wu, COO of ASE Group presented the packaging perspective. The question of margin and the acceptance of risk are critical issues.
“How do you construct a liability model that is based on the legal comprehension of the players, that is proportionate to the margin structure? How do you distribute the cost and liability across the supply chain based on economic principle? Today we have no leadership discussing those terms,” Wu stated.
Wu sees that the industry is in a “state of conflict and confusion” and must collaborate to survive the next evolution. “What has happened to the middle ground?” he mused. “There are no players left in the middle ground.”
September 1, 2009