In 1969, Busicom, a Japanese calculator company, hired Intel to design custom chips for their calculators. Intel engineer Ted Hoff proposed something different: instead of designing specialized chips for each function, build a general-purpose programmable chip. A computer on a single piece of silicon.
The Intel 4004 was born. Federico Faggin designed it. And then the patent question became: who owned it?
The ownership dispute
Busicom had contracted Intel to build the chips. Under work-for-hire doctrine, that made the invention Busicom's. Busicom held the original rights.
Intel's leadership, recognizing what they'd created, negotiated to buy back the rights in 1971 for $60,000 — returning Busicom's original payment. Busicom, struggling financially, agreed. Intel got the microprocessor patent for $60,000.
Within two years, Busicom went bankrupt. Intel became one of the most valuable companies in history.
What the patent actually claimed
US 3,821,715 covered the architecture of a general-purpose programmable computing device on a single chip. The claims were necessarily broad — this was genuinely new territory. No prior art existed because nothing like it had been built.
The patent was filed in 1973 and expired in 1990. For those 17 years (patent terms were measured from grant date until the Uruguay Round), Intel held foundational IP over the microprocessor concept.
The cross-licensing era
Rather than enforcing the patent to exclude competitors, Intel pursued a different strategy: licensing. They licensed the microprocessor architecture broadly, including to AMD (a decision Intel would later try to unwind), Texas Instruments, and others.
The rationale was market creation. A technology that only Intel made was a niche product. A technology that all the major chip companies made was an industry.
Cross-licensing became the template for semiconductors. Instead of patent fights, major players built portfolios and traded access. The result was an industry where the patents matter enormously — but primarily as negotiating chips (no pun intended), not enforcement tools.
When the patent expired
In 1990, when Intel's core microprocessor patents began expiring, the expected flood of competition didn't materialize — not because of other patents, but because of manufacturing advantage.
Intel had spent the 17-year patent window becoming the world's most efficient chip manufacturer. The fab technology, the process nodes, the yield improvements — these were Intel's real moat. The patent was the opening chapter. The manufacturing was the business.
This is the semiconductor pattern: patents protect the early period of a technology's commercialization, while manufacturing excellence protects the long-term business. The best practitioners use the patent window to build the capabilities that survive it.
What we learned
The microprocessor story established several principles that govern how technology patents work today:
Patents can be sold. The rights belong to whoever holds them, not whoever invented.
Licensing beats exclusion for infrastructure technologies. A standard that only one company uses isn't a standard.
Manufacturing moats outlast IP moats. The real advantage compounds in the fab, not the filing cabinet.
And $60,000 is sometimes the worst deal you've ever made — or the best, depending on which side of the table you sat.
PatentBrief is not a law firm. Nothing here is legal advice.