Bilski v. Kappos
The Supreme Court ruled that an application to patent a method for hedging risk in energy markets could not be patented, because it claimed nothing more than an abstract economic idea rather than a genuine invention.
The Court also rejected the Federal Circuit's rule that only inventions tied to a machine or that physically transform something can count as a patentable 'process,' leaving open, but unsettled, how future software and business-method patents should be judged.
How it got here: The patent examiner and the Patent Office's appeals board rejected the application; the Federal Circuit, sitting en banc, affirmed; the applicants asked the Supreme Court to review that ruling.
The Case in Depth
What happened
Two inventors applied for a patent on a method that would let energy buyers and sellers protect themselves against unpredictable price swings by locking in fixed rates with counter-parties who had the opposite risk exposure. The Patent Office examiner and its appeals board rejected the application, concluding it described only an abstract idea and mathematical relationships rather than a technological invention.
The question before the Court
Two energy traders wanted to patent a method for hedging against price swings. Could their idea be patented, and must every patentable process be tied to a machine or physically transform something?
Why it matters
The decision keeps the machine-or-transformation approach alive as one useful test, but not the only one, giving inventors of software and business methods more room to seek patents. At the same time, it confirms that basic economic concepts like hedging cannot be locked up by a single patent holder, so businesses can keep using well-known financial techniques freely.
What changes now
The Federal Circuit's rejection of the patent application is affirmed, so the application remains denied. The ruling is final on the merits, but the Court deliberately left the door open for the Federal Circuit and future cases to develop other tools, consistent with the patent statute, for deciding which inventions — especially software and other business-related methods — qualify as patentable processes going forward.
What this does not decide
The Court explicitly said it was not deciding whether software, medical diagnostic techniques, or other Information-Age inventions are or are not patentable, and it did not adopt any new comprehensive test for patentability beyond rejecting the two proposed categorical rules argued in this case.
Concurrences and dissents
Concurrence — Justice Stevens
Justice Stevens agreed the application should fail but would have decided the case on a narrower, historically grounded basis: that a general method of conducting business is not a 'process' under the Patent Act at all. He reviewed centuries of English and American patent practice concluding business methods were never meant to be patentable, and warned that the majority's broader reasoning about the ordinary meaning of 'process' could invite confusion.
Concurrence — Justice Breyer
Justice Breyer joined Justice Stevens' opinion in full, agreeing business methods are not patentable processes, but wrote separately to highlight four points of agreement shared across the fractured Court, including that Section 101 has limits, that the machine-or-transformation test is an important but non-exclusive clue, and that producing a 'useful, concrete, and tangible result' is not enough for patentability.
How the Court got there
The legal reasoning, step by step
- The Court first asked whether the 'machine-or-transformation test' — a rule asking whether an invention is tied to a particular machine or physically changes something — could be the exclusive way to decide whether an invention counts as a patentable 'process.' It concluded that courts cannot read limits into the patent statute that Congress did not write there.
- Because the statute's own definition of 'process' does not require a tie to a machine or a transformation, and the ordinary meaning of the words 'process, art or method' doesn't demand it either, the Court held the machine-or-transformation test remains a useful clue for spotting a patentable process but is not the only permissible test.
- The Court likewise rejected a categorical rule excluding all methods of doing business from patent eligibility, noting that the ordinary meaning of 'method' can include ways of conducting business, and that a separate patent-law provision creating a defense for people who already used a 'business method' before someone else patented it shows Congress assumed such patents could exist.
- Rather than adopt either broad categorical rule, the Court resolved the case using its long-standing precedents barring patents on 'abstract ideas' — a judge-made exception to the broad text of the patent statute recognized for roughly 150 years.
- Applying that rule, the Court found the application's core claims described nothing more than the basic economic concept of hedging risk, laid out as a series of steps in one claim and as a mathematical formula in another, with the remaining claims merely limiting that same idea to the energy market or adding routine statistical techniques.
- Because letting the applicants patent this idea would effectively give them control over the underlying concept of hedging risk in any field where it might be used, the Court concluded the application sought to patent an unpatentable abstract idea rather than a genuine patentable process.
Doctrinal impact
Cases affected by this decision
Reaffirms Gottschalk v. Benson (409 U.S. 63)
The Court relied on Benson's rule that transformation is a helpful clue but not a mandatory requirement for a patentable process.
Reaffirms Parker v. Flook (437 U.S. 584)
The Court used Flook's rule that adding minor extra steps cannot turn an unpatentable abstract idea into a patentable process.
Reaffirms Diamond v. Diehr (450 U.S. 175)
The Court relied on Diehr's rule that applying an abstract formula within a real industrial process can be patentable.
Limits Cochrane v. Deener (94 U.S. 780)
The Court said this older ruling's language was never meant to make transformation the only possible test for a process.