By Ed Korczynski, Senior Technical Editor, Solid State Technology
Possible collaborative solutions to bridge the looming R&D gap were discussed by Ho-Kyu Kang, representing the advanced technology development team within Samsung Semiconductor’s system LSI division, in a presentation at The ConFab.
Like IBM’s Calhoun, Kang emphasized that the investments needed to continue moving along the Moore’s Law trend-line call for substantial collaboration, including joint-evaluation and joint-development of new technologies, as well as cross-licensing and outsourcing of non-core-competencies. Also, industry-wide benchmarking and standardization is needed more than ever before.
The global semiconductor industry has an established history of collaboration, and so we can analyze past successes to extract lessons that may help us develop ever more complex manufacturing technologies. In general, any collaboration probably needs to be directed toward high-risk and correspondingly high-return investments. “We have seen that the more the risk, the greater the return,” Kang said. Also, with the need to establish mutual trust, it’s typically advantageous to work on multiple nodes over many years. There are many different specific business models that can work, so each unique combination of companies needs to negotiate mutually beneficial terms and conditions.
A common concern is whether collaboration will benefit not only the manufacturer but also the supplier. “I believe the supplier can benefit greatly, especially if the supplier is working with a ‘teaching’ customer, who can set mutually beneficial objectives and teach you your weakness, while helping develop your solution and the final bridge to manufacturing,” Kang said.
“Another question is whether there is any chance for collaboration with EDA vendors,” Kang noted. “I think there will be more opportunities for chipmakers to work with EDA vendors, and at a much deeper level than there has been until now, because of the new emphasis on DFM activities. So we have to invite EDA suppliers into the early stage of our new technology design and also into the early stage of our design build and definition of the project as a whole.”
However, the past also provides lessons that show the potential challenges of joint development program. Although there are obvious issues such as complicated handling of intellectual property (IP), and the extra time and cost to transfer the technology from the JV development site to each partner’s manufacturing line, many of the more basic challenges can be seen as more subtle. In particular, a lack of strong drive within the collaboration can result in lukewarm performance overall.
“Reasons for the lack of drive include the slow inflow of background IP because nobody wants to give away their own know-how,” Kang said. “So we have to some incentive for the inflow of background IP from the beginning. Also, in the course of doing development, many times there can be IP conflict among the partners,” he added. “Therefore, IP concerns need to be clearly addressed in the partnership contract.”
There are many other issues, including the residual need for partners to differentiate their technology. But all these problems can be solved, Kang said, by setting clear development goals and milestones combined with strongly structured communications can ensure that the collaboration stays on track. — E.K.