Patenting Quantum at the European Patent Office: when broad claims can become a risk

Suppose a quantum computer solves a fluid dynamics optimisation problem, outputting a bit string for which a customer pays good money.

Is the bit string itself a “product”? Most people might instinctively answer “no”. After all, it is just information. By contrast, your patent counsel would at least want to consider how the invention could be commercialised and whether the bit string might not in fact be the product.

And that brings me to an interesting challenge for quantum patenting before the European Patent Office (EPO).

What does “broad” mean in a quantum patent?

There are many reasons why broad patent claims are better than narrow patent claims, for example:

• harder for competitors to design around;
• commercially more flexible;
• capable of covering future business models that may not even exist when the application is filed.

Your patent counsel will naturally want to assist with all of the above, and more.

For quantum technologies, however, claims can sometimes sit close to the boundary between the abstract and the technical, and in such a scenario “broader” can easily translate to “abstract” (or “non-technical” in EPO vernacular). And that distinction can become important years after a patent has been granted.

For a slightly more detailed thought experiment, consider a quantum computer used to optimise an aerofoil design. A corresponding patent claim might conclude with:

[…] and obtaining a quantum state representative of an optimised solution for aerofoil design.

This is arguably a broad claim, concluding even before a measurement is made. The quantum computation itself is arguably where the real innovation occurs, so that can seem attractive. Moreover, patent attorneys would naturally favour this type of claim, particularly if the EPO examiner is prepared to allow it. And this is where things start to become interesting.

Patent grant is not the end of the story

A granted patent may later face attack before, say, the EPO Opposition Division or the Unified Patent Court (UPC). At that point, a challenger may argue that the claim lacks technical character because it merely defines something that could later be used for a physical purpose. That argument may be raised even if the EPO examiner was satisfied during examination, and indeed the challenger may be successful.

Amending the claim to introduce additional claim features tying the “abstract” quantum computing result to physical reality might seem like the best way forward. For example, adding the following steps:

• perform a measurement on the quantum state;

• generate design data from the measurement outcome; and

• manufacture an aerofoil using that design data.

This looks like an elegant solution. Problem solved?

The hidden trap

The amendment intended to save the patent, by making technical what was deemed abstract, may in fact have created an entirely new problem. According to the thought experiment, the amended claim now covers what is in essence a manufacturing process, concluding with manufacture of an aerofoil. According to European patent practice (see Article 64(2) of the European Patent Convention), a process directed to manufacture also protects products directly obtained by that process.

In other words, amendment can result in a process of manufacture and, by extension, adds protection for the products directly obtained. By contrast, the original claim was a method of operating a quantum computer, i.e. not a method of manufacturing and not protecting products.

The issue arises because the above amendment means that aerofoils are now protected after grant when they were not protected before grant. Extending the scope of protection after grant is a problem, because the rules of the game change with grant of a patent: the patentee is not allowed to extend the scope of protection following grant (see infamous Article 123(3) EPC). At least in a clear-cut manufacturing case, the Article 64(2) + Article 123(3) trap has sprung: introducing physicality into an abstract claim after grant has resulted in a process of manufacture, by virtue of introducing measurement and other “downstream” acts. That is, in simplified terms, the trap illustrated by recent EPO Board of Appeal decision T 0847/24.

So, what about scenarios that are not clear-cut manufacturing cases? A manufactured aerofoil is clearly a product whereas a quantum state is probably not. What about everything in between?

For example, what about:

• a data file containing instructions for a 3D printer to produce an aerofoil; or
• a bit string representative of a measurement outcome?

At what point does information become connected to physical reality such that the method becomes a process of manufacture and the outcome is a product?

Imagine a company paying for quantum compute time and receiving a bit string representative of the fluid dynamics optimisation problem. Arguably the bit string is the “product”, at least in terms of economic reality.

I am not aware of any decision by the EPO answering this question. And yet this question feels increasingly relevant as more value is created not by physical goods themselves, but by computational outputs.

Practical pointers

Broad claims remain incredibly valuable. However, when discussing pending EP patent applications with patent counsel (or, even better, already when drafting such applications), it may be worth asking:

• What happens if a granted claim is later characterised as too abstract?
• Do we have fallback positions involving measurement?
• Does commercial exploitation of the invention involve “manufacture” in some sense?
• Have we considered pre-grant claim amendments, for example introduction of a further claim?

Patent attorneys are trained to ask: how broad can we make this claim? For quantum technologies, there may be a second question worth asking: how close to physical reality should the claim end? Because if the broadest claim stops within the abstract, the amendment needed to save it after grant may no longer be available.