Amgen Inc. v. Sanofi
The Court unanimously struck down Amgen's broad patents on cholesterol-fighting antibodies, ruling that Amgen claimed ownership of millions of antibodies it had never actually shown scientists how to make.
The decision reinforces a bedrock rule of patent law — the broader the claim, the more an inventor must explain how to make the invention — with significant consequences for how pharmaceutical and biotech companies patent new drugs.
How it got here: The federal district court granted judgment for Sanofi finding the patent claims not enabled; the Federal Circuit affirmed; Amgen petitioned the Supreme Court, which agreed to hear the case.
The Case in Depth
What happened
Amgen and Sanofi are rival pharmaceutical companies that each developed drugs to treat high LDL ("bad") cholesterol by blocking a naturally occurring protein called PCSK9. In 2014, Amgen obtained two broad patents claiming exclusive rights to the entire family of antibodies that bind to PCSK9's "sweet spot" and prevent it from degrading the body's LDL receptors. Amgen described only 26 specific antibodies by their chemical makeup, but claimed rights over potentially millions more. When Amgen sued Sanofi for infringement, Sanofi argued the patents were invalid because Amgen had not actually taught scientists how to make all those antibodies.
The question before the Court
Can a pharmaceutical company hold a broad patent claiming millions of antibodies it hasn't taught scientists to make, simply by describing what those antibodies should do?
The Court's answer
No — the Court ruled unanimously that Amgen's patents were invalid because they claimed exclusive rights to millions of antibodies without adequately teaching scientists how to make most of them. Patent law's enablement requirement — written into the Patent Act since 1790 — demands that a patent's description be detailed enough that a skilled scientist can actually make and use everything the patent claims to own. Amgen identified only 26 specific antibodies by their chemical structure, yet its patents swept across potentially millions more.
For the remaining antibodies, Amgen offered only two general methods the Court found were really just research assignments: a "roadmap" directing scientists to generate and screen large numbers of candidate antibodies through trial and error, and a "conservative substitution" technique requiring scientists to tinker with known antibodies and test the results. Neither method reliably enabled a skilled scientist to make the full claimed class — they left researchers where Amgen found them, hunting through vast chemical space on their own. The Court's core principle: the more a patent claims, the more it must enable.
Curious how the Court got there? See the step-by-step legal reasoning →
Why it matters
Pharmaceutical and biotech companies that seek sweeping patents on entire categories of drugs — defined by what they do rather than how to make them — now face a firm reminder that they must actually teach scientists to make everything they claim. Companies that invested heavily in early-stage antibody research cannot lock up entire therapeutic fields without providing detailed, workable guidance to skilled scientists.
What changes now
With the Supreme Court's unanimous affirmance, Amgen's two broad genus patents are finally invalidated. Sanofi's competing cholesterol drug, Praluent, can remain on the market without fear of liability under those patents. Other pharmaceutical and biotech companies will need to revisit how they draft patent claims for broad classes of biological molecules, ensuring their specifications teach scientists how to make the full scope of what they claim — not just a handful of examples.
What this does not decide
The Court did not establish a separate, stricter enablement standard for "genus" patents. It also left open whether general research methods like the "roadmap" or "conservative substitution" could enable broader claims in other patents — for example, where the inventor has identified a shared quality common to every functional version of the invention.
How the Court got there
The legal reasoning, step by step
- The patent law 'enablement' requirement — found in 35 U.S.C. §112(a) — reflects the basic deal behind patents: an inventor gets a temporary monopoly in exchange for disclosing enough information that skilled scientists can actually make and use the invention once the patent expires. This quid pro quo has been in federal law since 1790.
- The Court traced three older precedents to establish the governing rule. In O'Reilly v. Morse (1854), the Court invalidated a claim that covered all means of telegraphic communication, not just what Morse had actually invented. In The Incandescent Lamp Patent (1895), the Court rejected a claim to all fibrous and textile materials for incandescent lighting when only carbonized paper had been shown to work. In Holland Furniture (1928), the Court invalidated a claim covering all starch glues that functioned like animal glue, because making any version required elaborate experimentation. Together, these cases establish the rule: if a patent claims an entire class of things, the specification must enable a skilled person to make and use the entire class.
- The Court acknowledged two limits on this rule. First, an inventor need not detail every possible version of an invention — a few examples may suffice if the specification identifies a shared quality running through the entire claimed class that a skilled person can reliably apply. Second, some degree of testing or adaptation by skilled practitioners is acceptable, so long as it is reasonable given the state of the field.
- Turning to Amgen's claims, the Court found they swept far beyond the 26 antibodies Amgen had identified by amino acid sequence. The claims covered every antibody that binds to specific parts of PCSK9 and blocks it from degrading LDL receptors — a 'vast' class potentially numbering in the millions. Amgen, like Morse and Sawyer-and-Man before it, sought 'sovereignty over an entire kingdom' defined only by the function it wanted to monopolize.
- Amgen's two offered methods — the 'roadmap' and 'conservative substitution' — did not solve the problem. The roadmap merely described Amgen's own trial-and-error discovery process: generate lots of antibodies, test them, see what sticks. Conservative substitution instructed scientists to swap amino acids in known antibodies and retest — an equally uncertain approach given that even a single amino acid change can unpredictably alter an antibody's behavior. Both methods left skilled scientists to engage in open-ended experimentation to find what works — more a 'hunting license' than an enabling disclosure.
- The Court also rejected Amgen's policy argument that a ruling against it would destroy incentives for breakthrough inventions. Balancing incentives for inventors against the public's right to benefit from disclosed knowledge is a judgment for Congress, which has maintained the enablement requirement since 1790. The Court's job is to enforce that requirement as written, not to relax it based on a patent holder's view of sound innovation policy.
Doctrinal impact
Cases affected by this decision
Reaffirms O'Reilly v. Morse (15 How. 62)
Reaffirmed that a patent cannot claim all means of achieving an end when the inventor has not shown how to make them all.
Reaffirms The Incandescent Lamp Patent (159 U. S. 465)
Reaffirmed that a broad functional claim fails unless the inventor identifies a quality shared across all claimed embodiments.
Reaffirms Holland Furniture Co. v. Perkins Glue Co. (277 U. S. 245)
Reaffirmed that claiming a functional class requiring elaborate experimentation to identify valid members fails the enablement requirement.