IBM recently announced that they had signed a definitive agreement which will see GlobalFoundries (GF) acquire IBM’s global commercial semiconductor technology business, including intellectual property, world-class technologists, and technologies related to IBM Microelectronics. IBM and GF have a history of collaborating on technology development. From participation in the Common Platform Alliance, to the more recent announcement that IBM was sending up to 200 employees from their East Fishkill R&D facility to GF’s R&D center in Malta, NY. Not surprisingly, GF was widely speculated to be the one chosen to take over IBM’s semiconductor operations.
We thought we would take a look and see what intellectual property would likely be part of this deal. Patenting is an important barometer of innovation according to IBM, and they have topped the annual list of US patent recipients for over twenty years. Using our patent analytical techniques here at Chipworks, we scanned over 42,000 IBM active US granted patents to identify those which our classification methods determined were related to semiconductor technology. The result was a list of approximately 7,300 patents. We decided to exclude about 800 patents, which were related to packaging technology, since it is reported that IBM is keeping its packaging operation in Bromont, Quebec. We also assumed circuit-related patents would remain with IBM. We then looked at US grants which have at least one foreign equivalent in active semiconductor countries and found approximately 2,100 patent families. So, we can quickly identify about 9,400 US and foreign patents that could potentially be part of the deal. The figure of 10,000 patents reported in EETimes may be reasonably accurate.
Click here to view table.
Regular readers of our blogs are probably familiar with patent landscapes. These are graphical representations of portfolios that cluster patents according to concept. Each dot on the landscape represents a single patent, and a dense group of dots on the landscape is one indication of strength in a particular technology area. Areas of high density have keywords attached to provide context.
In the landscape presented here, we plotted all GF US granted patents in red (about 5,000 of them) and overlaid the 9,400 IBM semiconductor patents identified above in blue. Our first observation is that the IBM patents complement GF’s IP in core semiconductor technology. There is good representation from both IBM and GF in the regions labeled with keywordsgate/region/layer and layer/metal/dielectric. Patents related to core CMOS technology reside in these clusters. Clusters located around the periphery of the landscape with keywords such as waveguide, ferromagnetic, nanotube, etc., relate to more focused technology that IBM may or may not transfer. Finally, the large cluster of GF patents located at the upper right of the landscape are predominantly circuit related. We made the assumption that IBM would likely retain their circuit IP, hence the noticeable absence of IBM in this region.
We colored IBM patents that have at least one foreign equivalent in light blue to determine if there are any areas which IBM considered more valuable. The ratio of US to foreign patents for IBM is lower than what we typically observe for other technology companies, so we think this is one method to identify what they consider to be their strong patents. With the exception of the nanowire cluster found at the lower left of the landscape, all areas appear to be well covered outside of the US.
In summary, the acquisition of IBM semiconductor will strengthen GlobalFoundries IP position, potentially give them access to next-generation niche technologies, a couple of leading-edge fabrication facilities, plus a $1.5B cash bonus. This is the point where we gauge reaction of the markets to the news, to determine if this is a good deal, but since GF is privately held, we can only assume that this is a win-win for both companies.