patent valuation: don’t forget the technology
Patent valuation is the holy grail of IP. Everyone wants to do it and many claim they can, but it is actually quite difficult. The quest to valuation is more entailed than a single “click” because it relies on knowledge of technology. Ignoring technology is at the valuator’s peril.
The patent analytics vendors represent today’s baseline. They make valuation look easy and are often seen as the go to valuation source. Simply move the cursor over a “flag” in a patent landscape and a value for that patent appears. It is quick and it appears convincing because the process outputs a very specific number. Right or wrong the number sticks in your head.
Generally, they use a large equation that weights all sorts of data associated with a patent. This might include the number of citations, the word count of the Description, the word count of the Claims, the Inventor, filing trends in Classification Codes etc. It is all very complicated, but the power of data makes quick work of the task and a value pops out. But, is valuation that simple? Is the value all there in the data, ready for extraction?
In the summer of 2022 I consumed marketing literature from one of these patent analytics vendors. More specifically, it was a brochure discussing how their services provide insights into the semiconductor industry. Nice, a field I know quite well. One table presented a well known integrated semiconductor manufacturer’s five most valuable patents. The second patent in this table was assigned a very precise value of $11,970,000. The table also indicated the patent would be alive for two years from that time i.e. it would expire in 2024.
Let’s go back a few steps and remember that a patent discloses and legally protects an aspect of technology. After simply reading the patent’s title I felt uneasy with the valuation because it suggested an area of technology that was, maybe, only going into production. Even then, use of the patented aspect of technology is not guaranteed and the patent was said to expire in two years. The valuation might be on shaky ground.
It is time to take a closer look.
Field Effect Transistors (FETs) are the most fundamental component of Integrated Circuits (ICs). Today’s ICs are mostly fabricated with either planar or finFET technology, where finFETs are at the cutting edge, being found in ICs requiring the highest performance. These are the processors in your laptop and System on Chips (SoCs) in your phone. Intel was the first to commercially produce finFET-based ICs in 2011, making the switch from planar FETs to finFETs at their 22 nm process node. Others squeezed a bit more from planar FETs, extending their life to between 20 and 16 nm, depending on the particular flavour. FinFETs continued “scaling” or shrinking to 5 nm.
Some manufacturers dropped out of the “nm” race along the way as it simply became too expensive. TSMC and Samsung are two that remain. Samsung issued a Press Release in late June 2022 announcing commercial production of 3 nm Gate All Around (GAA) FETs i.e. the next generation of FET. As September 2023 approached TSMC remained at 5 nm finFETs.
On September 12th 2023 Apple held its yearly iPhone Event. It is Apple’s showcase for their newest A-series processor. Prior to the 12th it was widely believed that the 5 nm node will be the final one for TSMC finFETs and they too would move to 3 nm GAAFET technology.
On the 12th Apple unveiled the iPhone 15 Pro and the A17 processor. We were told the A17 is made with a 3 nm process, but there was no mention of GAA. A few rumblings from colleagues suggested TSMC stuck with finFETs for 3 nm. OK the case for the valuation is looking shakier. We waited and waited for the reverse engineering houses to get their hands on the A17 and do their thing. After a bit more waiting there it was (see below). The A17 is manufactured with TSMC’s 3 nm finFET process. In one fell swoop, TSMC would not likely move to GAA until after the patent is said to expire.
A bit of reading beyond the Title confirmed the patent is related to GAAFETs. There are generally two genres of GAAFET; nano-wires and nano-sheets. Samsung’s Press Release indicated they adopted the latter.
Remember, a patent discloses a very particular aspect of technology. In an area like semiconductors these aspects can be quite fine. In this case the patent presents a nano-wire finFET structure. Further, the Claims refer to a horizontal nanowire channel and details of SWS geometry and orientation with respect to surrounding structure. All of these are important details if the owner wants to extract licensing revenue from the patent.
Does it look like Samsung might use the patented technology? And TSMC? This is where I will end the analysis. However, I will say that the stated valuation might just be on the optimistic side.
This is not an isolated case. In fact, it jumped out at me because it was highlighted in a brochure claiming expertise in an area. “OK since you are pointing out this patent, let’s look a bit closer”. In 2019 I wrote about IBM’s so-called coffee drone patent. IBM, AI, word count, filing date, citations; all sorts of data likely point to a high valuation. But, it is important to actually look at the technology. In 2020 I wrote about the concept of IP-rich. Again, the concept of being IP-rich is a function of the underlying technology because that is the fundamental basis of the patent.
Government, investors and society in general are fascinated with big data and the ability to “crunch” it. For sure big data provides insights that were not previously possible. However, in the case of patent valuation, at least, the underlying fundamentals i.e. technology can not be ignored. Sometimes all of the patent’s data points in one direction, but the technology is simply not adopted. Then, the patent protects technology that is not made, marketed or sold and valuation suffers, to say the least. Not appreciating or considering the technology is perilous.