The United States grants hundreds of thousands of patents a year. A small minority ever appear in a product, and an even smaller minority generate revenue exceeding the cost of obtaining and maintaining them. That is not a failure of the system so much as a feature of it: a patent is a legal right to exclude, not a business, and the distance between the two is where most inventions stop.
What a patent actually gives you
A granted patent confers the right to prevent others from making, using or selling the claimed invention for a limited term. It does not confer the right to practise the invention, which may be blocked by other patents, nor does it certify commercial value or even that the thing works at scale. Examination tests novelty, non-obviousness and adequate disclosure — not manufacturability. Grant statistics, examination procedure and the full searchable corpus are published by the United States Patent and Trademark Office.
Read moreWhy AI Data Centers Are Reshaping U.S. Power
Claim drafting determines commercial worth more than the underlying science does. Broad claims that are hard to design around are valuable and hard to obtain; narrow claims are easy to obtain and easy to avoid. Most of the money in patent practice is spent on this distinction.
From laboratory to pilot: the valley of death
Between a working prototype and a manufacturable product sits a stage that is too applied for research funding and too unproven for commercial capital. Scaling a chemical process from grams to tonnes, qualifying a component for automotive reliability standards, or moving a device from hand assembly to tooled production are engineering problems as hard as the original invention and far less glamorous.
Public programmes exist specifically to bridge this gap, including translational research funding and small-business innovation grants coordinated across agencies, with substantial programmes at the National Science Foundation and the national laboratories of the U.S. Department of Energy. University technology transfer offices, empowered by legislation allowing institutions to own inventions made with federal funding, are the other main conduit.
Read moreWhat a 10-K Really Tells You About a Company
Licence or build
An inventor with a granted patent faces a fork. Licensing transfers commercialisation risk to a company with manufacturing and distribution already in place, in exchange for a royalty typically expressed as a low single-digit percentage of net sales. Building a company retains far more of the upside and requires capital, a team and years.
The choice is usually made by the nature of the invention rather than the ambition of the inventor. An improvement to an existing product category is best licensed to an incumbent, because the incumbent’s channel is the asset. A genuinely new category has no incumbent to license to, which forces company formation.
Enforcement is a business decision
A patent is only as strong as the willingness to litigate, and litigation in the United States is expensive enough that the threat is credible mainly for well-capitalised holders. Administrative review procedures allow a challenger to attack validity at lower cost than a full trial, and a significant share of challenged claims are cancelled or amended — which means a granted patent should be treated as a presumption of validity rather than a guarantee.
Read moreThe Evidence on Remote Work and Productivity
International protection multiplies both cost and complexity, since patents are territorial. Coordinated filing through the international application system administered by the World Intellectual Property Organization defers national costs but does not eliminate them, and maintenance fees over a full term across major markets run well into six figures.
The practical sequence
Successful commercialisation tends to follow a recognisable order: establish that a customer will pay before spending heavily on prosecution; file early enough to preserve rights but late enough to know what the claims should cover; solve manufacturability before scaling marketing; and treat the patent as one element of a defensive position that also includes trade secrets, data, brand and speed. Inventions protected only by a patent are rarely protected for long.
Why most patents never become anything
A granted patent is a right to exclude, not a product and not a business. The overwhelming majority are never commercialised, and the reasons are rarely technical. A patent may cover a step that is necessary but not sufficient, requiring licences to other patents held by other parties before anything can be built. It may protect a process that works at laboratory scale and fails at manufacturing scale, where yield, throughput and cost per unit are the binding constraints. It may simply have been filed defensively, to prevent a competitor from claiming the territory, with no intention of practice.
Grant, assignment and maintenance-fee records published by the U.S. Patent and Trademark Office make the attrition legible, since abandonment through unpaid maintenance fees is itself a verdict on commercial value. The path that does work usually involves the same unglamorous sequence: reduction to practice at pilot scale, a manufacturing partner willing to invest in process engineering, regulatory clearance where applicable, and a customer who was going to buy something anyway. The invention is the beginning of that sequence rather than the substance of it.
