The Top Ten Patent Wars – Semiconductors (#7)
The War That Never Happened
There has been so much patent litigation regarding semiconductors over the past 70 years or so that even a brief synopsis would exceed one post. Instead, let’s take a look at the war that the semiconductor industry expected, but never happened – the Semiconductor Chip Protection Act (SCPA).
Semiconductors Have Always Been the Subject of Patent Wars
Highly complex semiconductors are the basis for much of modern electronics technology, powering nearly every complex computing device in the world. They work by conducting electrical current at certain points, and insulating at other points, hence as a “semiconductor.”
Patent protection was obviously the type of protection most-sought for these types of inventions, starting with the transistor that was invented by William Shockley and others in the late 1940s. For example, there was a significant patent dispute between Fairchild Semiconductor and Texas Instruments in the late 1950s and early 1960s about who originally invented an improvement over the transistor, the modern integrated circuit (it turned out they both did, a few months apart). These increasingly sophisticated integrated circuits (aka “ICs” or “computer chips”) later become commonplace throughout what became known as Silicon Valley and beyond (because the chips were initially made out of “silicon”).
The Industry Sees Another War Brewing
At one point, in the 1970s, the semiconductor industry believed that an IP war was coming because the IC designs had become so intricate that chip “pirates” were buying chips that companies had spent years developing, reverse engineering them and then selling them at a fraction of the price. Essentially, they were counterfeiting the chips.
This infringement fell outside the realm of patent because these pirated designs were not inventions, but rather designs etched onto the chips of already patented devices. But they also fell outside the scope of copyright protection, because these designs were functional and not aesthetic, as they designed to control the conductivity of the products they powered. And trade secret protection obviously failed as well once the design was introduced to the public. After much lobbying by the semiconductor industry, Congress settled on creating a new type of copyright for a “mask work” for these semiconductor designs. In 1984, Congress passed the Semiconductor Chip Protection Act and created an entirely new IP right, the first new IP right in 100 years. There is an entire chapter in the copyright statutes protecting the designs of these semiconductors, referring to such IP protection for “mask works”. See 17 USC, Chapter 900.
Yet, according to a study from 2012, there were only a little over 1000 mask works registered with the Copyright Office between 2008 and 2012. (See “Copyright Registrations: Who, What, When, Where and Why”, Oliar, Patterson, Powell, Texas Law Review, Vol. 92:1225). Only one federal case, Brooktree Corp. v. Advanced Micro Devices, Inc., 977 F.2d 1555 (Fed. Cir. 1992), ever analyzed the statute in detail and found a violation. Brooktree did upheld a substantial jury verdict under the SCPA, but only where there was a clear ability to compare the protected work and the infringing chip. Mask works have just become a footnote to copyright law, and essentially irrelevant to a ubiquitous modern technology. So what happened?
Moore’s Law Created a Cease Fire
Well, what happened is Moore’s law, both literally and figuratively. Moore’s Law, famously named after Gordon Moore, the co-founder of Fairchild and Intel, who in 1965 theorized that the number of components in a chip would double every year (later revised to two years). In other words, technology goes fast. It went too fast for the SCPA to launch any sort of IP war for this new IP right. In alignment with Moore’s Law, semiconductor chip design has advanced at such a fast pace that the SCPA turned out to be superfluous. Semiconductor chip design is so complicated that it would be impossible to pirate a design, essentially. For example, as of 2017, one AMD Epyc server product used about 19 billion transistors on a single chip.
By the time a “pirate” would be able buy a chip publicly, and to reproduce the design of, let’s say, 19 billion different transistors, it would be obsolete and protection under the SCPA would be unnecessary. And even pirating a small portion of the design would be useless because the chips are so specialized to be used for a specific company or product that such a design would not be commercially viable.
SCPA Unlikely to Ever Be Relevant
The IP war feared by the semiconductor industry failed to materialize. Semiconductor inventions are still protected by patents and have been litigated extensively over the years. Any such “piracy” that concerned the industry would probably fall under trade secret law, given that most companies choose to keep designs and aspects of their processes secret. Companies do not choose to disclose their designs to the public under the SCPA, which is why there are so few mask work registrations. The SCPA appears to be destined for the dustbin of history.