In 1994 Monsanto was granted a patent for genetic material from a virus that it used to incorporate new genetic material into plants. This enabled Monsanto to go forward and patent a gene that makes seeds resistant to the chemicals farmers use on weeds. Monsanto now believes its patents cover not only its seeds, but every single plant or seed created from the original seeds. On the other side of the argument, farmers believe that once they purchase seeds from Monsanto, they own the seeds outright and every crop grown from the original seeds thereafter.
In 2013, a quiet ruling in Bowman v. Monsanto Co. settled a narrow dispute about soybeans. It also fixed a broader principle: owning a thing is not the same as owning the right to reproduce what is inside it. More than a decade on, that line runs through agriculture, vaccines and artificial intelligence.
It began with a farmer buying soybeans.
Vernon Bowman did not purchase specialist seed from a manufacturer. He bought commodity beans from a grain elevator, planted them, sprayed herbicide, and saved what survived. In earlier eras, that sequence would have been unremarkable. By the time the case reached the Supreme Court of the United States, it had become a test of something else entirely: whether a purchaser may use a self-replicating product to make more of a patented invention.
The Court’s answer was concise. You may use what you have bought. You may not use it to make new copies of the patented invention. Planting seed for consumption is use; planting seed to create further seed carrying the same patented trait is making.
That distinction, between use and reproduction, now sits quietly beneath a great deal of modern technology.
The narrow case that became a general rule
Monsanto’s claim was not that it owned soybeans in the abstract. It asserted rights over a specific engineered trait that makes plants resistant to glyphosate. Bowman’s method of buying mixed beans, selecting the resistant plants, and replanting was, in the Court’s view, a deliberate way of reproducing that patented trait without paying for it.
The doctrine the Justices relied upon, patent exhaustion, is usually straightforward. Once a patented item is sold, the seller cannot control that particular item. The difficulty here was biological. Seeds do not remain as they are. They multiply. If exhaustion extended to all future generations, one sale would effectively end the patent. The Court declined to go that far.
It instead fixed a boundary that has proved durable: ownership of the thing does not include the right to reproduce the invention embedded within it.
Agriculture after Bowman
For farmers, the practical effect is familiar. Commercial seed is commonly sold under licence. Saving and replanting that seed, where it carries patented traits, requires permission. The ruling did not create that system, but it cemented its legal footing.
The harder questions sit just beyond the field.
Is it reasonable that a biological process of growth and reproduction can be treated as the making of a protected article? The Court answered in the affirmative, largely for economic reasons. If the first sale enabled unlimited copying, the patent would have little value. That logic is not controversial in machinery or chemistry. It is more contested when the “machine” is a living organism.
There is also the matter of scope. Modern crop traits are covered by layered patents: genes, methods of insertion, uses, and sometimes combinations with particular treatments. Designing a clean alternative that achieves the same end without touching those claims is possible in principle and difficult in practice. It is, nonetheless, the route the law encourages.
From fields to laboratories
The soybean dispute looks less parochial when set against biotechnology.
Many contemporary vaccines and therapies depend on processes that replicate biological material: viral vectors grown in culture, cell lines that expand to produce a compound, genetic constructs that are copied in manufacturing. The Bowman line – use versus reproduction – maps onto these processes with little adjustment.
Possession of a vial does not confer the right to scale it into production. Manufacturing requires permission because it involves making new instances of a protected invention. During the pandemic, when calls for wider manufacturing met the realities of patent rights, that distinction moved from legal theory into public argument.
Courts have also drawn a boundary between discovery and invention. In Association for Molecular Pathology v. Myriad Genetics, the same Court held that naturally occurring DNA is not patentable, while synthetic constructs may be. Together, these decisions sketch a consistent, if contested, position: nature itself is not owned, but human-made applications of it may be, and reproducing those applications remains controlled.
The same logic in software and AI
The rule appears again, more quietly, in software.
A program is easy to copy. The law nonetheless treats copying as making a new instance of the protected work. Buying software permits use; it does not permit duplication beyond the licence. That is familiar.
What is less settled is the status of systems that generate new outputs, most obviously, artificial intelligence. Tools such as Suno do not reproduce a single work; they produce new compositions from learned patterns. The legal debate there sits in copyright rather than patent, but the underlying question is recognisable: when does producing something new amount to copying something protected?
Bowman does not answer that question directly. It does, however, establish the judiciary’s willingness to treat replication—however natural the mechanism—as an act of making when it recreates a protected invention.
The uncomfortable middle ground
There are two competing intuitions, and both have force.
One holds that ownership should be meaningful. If a person buys a thing, the instinct is that they should be able to use it fully, including in ways that follow naturally from what it is. Seeds grow. Code copies. Cells divide.
The other holds that invention requires protection. Without a period of control over copying, the argument runs, there is little incentive to invest in developing new traits, therapies or systems.
The Court chose the second intuition where the two collide. It did so in a narrow case, with a farmer who plainly intended to replicate a patented trait. The principle it articulated has proved broad.
What has changed and what has not
Since 2013, the landscape has shifted in scale rather than in kind.
Seed markets have consolidated further, though new entrants and public breeding programmes continue to operate. “Open seed” initiatives have grown, offering varieties that may be saved and shared without restrictive terms. In biotechnology, the tension between intellectual property and access has become a standing feature of policy debates. In software, licensing models have diversified, from proprietary control to open source, without altering the basic rule that copying is controlled.
What has not changed is the boundary Bowman set. The buyer owns the instance. The right to reproduce the invention remains with the patent holder.
The question that remains
The case did not settle the moral argument. It clarified the legal one.
Where should the line be drawn between shared natural processes and time-limited control over human-made applications of those processes? Reasonable people answer that differently. Legislatures adjust it; courts interpret it; markets respond to it.
For now, the soybean case stands as a quiet pivot point. A seed may still be planted. What cannot be assumed is that planting carries with it the right to make, and keep making, whatever a patent has been placed inside it.
That is a small distinction in wording, and a large one in consequence.
NOTES:
Listen to Bowman vs. Monsanto or read the transcript.
