A technical lead is worth what you can prove. Being first, or fastest, or best matters only if the evidence behind the claim survives a hostile reading, whether from a court, an examiner, an acquirer’s due diligence team or a competitor’s expert. That is where a lot of patent strategy quietly fails, and it is the lesson in an unlikely place: a quantum computing result.

The result was reported at the end of August. IBM and the University of Chicago ran 70 error-corrected logical qubits through a sampling problem in about 15 minutes, a task the team says would take leading classical methods an impractical length of time. The distinction from earlier claims is in the second half of the announcement. The circuit was designed so that errors could be detected during the run, giving statistical evidence that the answer was trustworthy, not just fast. Quantum results have been announced before. Verifiable ones are new.

Even so, the claim was challenged within days. Other researchers questioned how far the result could be pushed and argued that advantage is an ongoing contest rather than a milestone. That is the point. The strongest evidence a lab can produce still gets tested, and a company that has only its own confidence to offer will not last the first round.

Patent strategy runs on evidence, not on being first

The same test applies to any business that trades on a technical edge. The questions it will face are predictable:

– Does the patent specification actually teach the invention, with the best method the company knew at filing? Both are grounds of invalidity, and both are decided on what was written down at the time.
– Can the data behind the claimed result be reproduced, and is there a record of who did what, when, with which materials and code?
– If the edge is kept as a trade secret, can the company show what the secret is, that it was treated as confidential, and that the departing engineer took it (and, or course defend against lack of Best Method attack)?
– In a sale or a capital raise, will the technical claims in the information memorandum hold when the buyer’s advisers ask for the underlying records?

None of this is generated on the day it is needed. When Your Patent Isn’t Built for the Road Ahead shows an Australian patent failing on insufficiency and best method because the specification could not carry the weight placed on it. Built Before the Fight: What May’s IP Decisions Reward makes the wider point that disputes are decided by records made years earlier. And Where Is Your Leverage? What a Week of IP Disputes Teaches About Strategy argues for stress-testing patents and documenting data provenance before anything is asserted.

The IBM team built verification into the experiment rather than bolting it on afterwards. That is the model. Treat the proof as part of the invention, budget for it, and make someone responsible for it. A lead you cannot prove is a lead you do not own.

Read the article: IBM quantum computer solves classically intractable problem in 15 minutes, ScienceDaily, from materials provided by IBM and the University of Chicago.