Paul Boldt attempts to unravel Rambus' Acquisition of Unity Semiconductor
listen to this story
Where were we? Oh yes…
Of all the possibilities for Unity Semiconductor as of February 5th, 2012 I would have placed a sale to Rambus pretty low on the list.
Let’s break the discussion down according to timeline; the past, the present and the future. In this case the past will encompass 2011, the present will be the time around the acquisition, and the future will be the time going forward from the acquisition.
The past. Looking back there were really two timeframes that seemed to set a positive tone for 2011. The first was the third week of January. Press releases came out on the 18th and 20th of January, 2011 informing us of a new executive team, and a joint development agreement with Micron, respectively. In the latter release Mr. Eggleston is quoted:
“We are excited to be working with Micron, jointly developing CMOx™ in their state‐of‐the‐art facilities with their world‐class engineers. Adoption of new memory technologies is essential to the continuing evolution of new and improved computing and mobile devices.”
While not specifically in the above releases it is confirmed in later material that Micron invested in Unity.
Fast forward to the fall 2011, October 25th to be exact. Unity issues a press release entitled “Unity Semiconductor Solves Cross Point Memory Design Puzzle”. Again a quote from the release:
“Unity’s new Memory Tile circuit design, combined with our novel CMOx™ memory cell and vertical cross point architecture, are major steps towards the commercialization of Terabit (Tb) non volatile memories.” … “Unity has crafted the comprehensive set of design, process and device solutions required by manufacturers for implementation of our technology in leading edge semiconductor memory fabs.”
The press release reiterates a “broad licensing program”. Okay, the Micron partnership appears to be bearing fruit. I would however suggest it is was in fact a pivotal year for Unity.
So 2011 seems to be progressing well, but there are only roughly 3 months between the above October release and being acquired in early February. Did Unity already have feelers out at the time of the release? One might also speculate the due diligence was started before the end of 2011. Assuming this is a reasonable timeframe there was hardly time for the memory tile circuit design development to flesh itself out. Was any new circuit still in design / verification, was it taped-out, or was there working silicon at the end of October? If the circuit was still in the design stage they may not have had working silicon by the time of sale. Else, there may have been problems in the silicon that had just come back, raising concerns about cash runway. It does not seem that all the pieces were in place in late 2011.
The future. Cutting right to the chase, the most value from this deal will probably only be realized if Rambus advances the technology to a viable commercial state. On the face of it this looks like what they intend to do, as suggested in the following excerpt from the Rambus’ news release on February 6th:
“As part of this acquisition, the Unity team members have joined Rambus to continue developing innovations and solutions for next-generation non-volatile memory … With the addition of Unity, we can develop non-volatile memory solutions that will advance semiconductor scaling beyond the limits of today’s NAND technology. This will enable new memory architectures that help meet ever-increasing consumer demands.”
In the end Rambus may not want to market its own line of CMOx memory devices, but the most value from any licensing program will only be realized if they bring CMOx to market. There appear to be at least two considerations. First, over the last few years it seemed to be Unity that was the strongest advocate for CMOx. This is not to say that others were not researching in the area, but it is probably safe to say that Unity was a major driving force. Further, it is not clear if others will develop the technology as diligently as they might have knowing that a licensing entity is now in control of the Unity IP. Second, while not having read Unity’s IP portfolio there is a good chance many of the patent’s claims will at best, if not only, read on CMOx. So, we are back to the need to bring the technology to market.
Completely apart from how near or far the technology is from commercialization, it still seems weird that as one of the investors, Micron would have received money from Rambus. Business does indeed make strange bedfellows.