Everyone today understands the essential importance of third-party semiconductor intellectual property (IP) to the health of the semiconductor industry. Considerably less attention is given to the importance of a strong, serviceable portfolio of internal IP within a semiconductor company.
Since the beginning of the semiconductor industry, chip companies have had strong, thriving portfolios of IP that differentiated them from their competitors. Until the mid-1980’s, these companies were “vertically integrated” with their own foundries, EDA tools, and IP. That all changed in the 1980’s when TSMC came into the picture and began an era of disaggregation. No longer did a company need to have its own expensive foundry for producing chips. The same happened with EDA, where legions of companies such as Synopsys, Mentor, and Cadence came onto the scene to increase efficiencies even further. Companies could then focus on their true competency—design.
The final step of disaggregation began in the 1990’s, when third-party IP companies like ARM and Artisan appeared on the market to allow new semiconductor companies to form and build thriving businesses using third-party foundries, EDA, and now IP. These new-breed semiconductor companies (aka “fabless”) continued to differentiate themselves with their own “special sauce”—internal IP coupled with commodity, off-the-shelf IP, assembled in unique ways. Today, the most advanced chip companies closely guard their special sauce recipes by investing significantly in internal IP efforts to stay ahead of their competitors.
We begin here to discuss what semiconductor companies should be doing to maximize the value of their internal IP. But first, we have to define some terms, because all IP is not created equal. In general, there are two types of IP: reusable and salvageable. Reusable IP, as its name suggests, is explicitly designed to be reused as a package by someone other than the original designer. It is designed, documented, and verified with this in mind. Creating reusable IP is expensive and is typically relegated to designs that will be used many times (without modification) over its lifetime, which could be ten to twenty years in the case of synthesizable IP. Salvageable IP, on the other hand, is simply a design than can be repurposed from its original intent. It often lacks the formal documentation and independent verification methods that are needed for reusable IP. While the term caveat emptor is certainly applicable to engineers desiring to use salvageable IP, it remains the most common form of design reuse within semiconductor companies. Engineers simply must be aware that the salvaged IP needs to be re-verified in the context of the design they are working on and that the expectation is that the code may help accelerate their design, but the risk is fully on their shoulders.
Now that we have defined the two types of IP, we can address value. When we talk about the value of IP, we have to consider different contexts:
- The value of the IP within your company
- The value of the IP outside your company
- The tangible vs. strategic value of the IP
The value of the IP within your company is essentially the monetary value of having a piece of IP on hand for your engineering team to use as they do their work. The alternatives are to design the IP from scratch or purchase it externally. If the decision is to design, then a choice must be made to either design it as reusable IP or for a one-time, salvageable use. Obviously reusable IP has higher value than salvageable IP, but the reusability has a cost that is not difficult to calculate. Typically, if an IP can be reused without modification five times in its lifetime, it is worth the additional cost to make it reusable. The Global Semiconductor Alliance’s IP ROI Calculator tool can be useful for companies trying to understand and assess this.
The value of the IP outside your company is, quite simply, the monetary value of the IP to engineers outside of the company—in short, what others would be willing to pay to gain access to it.
Finally, we have the tangible value vs. strategic value aspect of IP. This is a bit more difficult to put a figure on, but in essence, this is an assessment of whether an IP may be a strategic differentiator to the company’s product, versus merely being related to improved efficiencies of their operation. Sometimes, the value of an entire product line may come down to a particular IP that a company possesses.
Oftentimes, the true value of an IP is trapped behind barriers that the semiconductor company has inadvertently erected for itself. The first barrier is related to the reuse vs. salvage state of the IP. If a salvageable IP has the capacity to be reused many times, then not converting it to be fully reusable represents a waste of the chip designers’ time every time it is used. The second barrier is related to awareness. One of the companies we’ve worked with at IPextreme related to me that they discovered that their company had amassed 32 different UART’s, developed over different times by different divisions. This happened simply as a result of the company lacking any organized system for capturing internal IP so that other engineers could possibly salvage previous work (not to mention trim some fat from the bottom line). The third barrier is familiar to many engineers—fear of sharing. It is a common fear that if my IP is shared, then I will be responsible for maintaining it, fixing bugs, etc. This is why companies should consider the organizational aspects of internal IP reuse programs. From our experience, the best companies have formal IP design teams responsible for creating and maintaining reusable IP and have other systems in place for storing salvageable IP that can be tapped on an “as is” basis.
There are a number of best practices that can be established for maximizing the value of internal IP. These are simple, but require executive level buy-in for success. In a nutshell, these are as follows:
- Having methodologies and policies that are endorsed at the highest level to facilitate IP reuse and salvage across the company.
- Creating awareness of what IP is available to engineers, both from internal and external sources.
- Planning at the product engineering level that anticipates roadmaps of needed IP, thus allowing sufficient time to decide whether IP should be purchased, if reusable designs ought to be planned, or if the plan is to salvage.
Some excellent videos have been produced on this topic and are available on IPextreme’s IP WATCH channel. Here you will find Warren Savage speaking at a GSA event in January 2013 on this same topic, as well as an excellent presentation from Kevin Klein of Freescale Semiconductor on IP reuse within a large semiconductor company.