Trade Secret

Trade secret is a category of intellectual property protecting information that has commercial value specifically because it is kept confidential, rather than because a government office has granted an exclusive right in it. In U.S. law, a trade secret is any business, technical, scientific, or engineering information — a formula, process, method, compilation, prototype, or design — that (1) derives independent economic value from not being generally known or readily ascertainable by others who could profit from it, and (2) is the subject of reasonable efforts by its owner to keep it secret. Both elements are load-bearing: information that is valuable but poorly guarded, or carefully guarded but not actually valuable, does not qualify.

Unlike a patent, copyright, or trademark, a trade secret is not registered with any government office and carries no fixed term — protection lasts exactly as long as the information stays secret and continues to meet both statutory elements, and ends the moment either one fails (independent discovery, publication, or a lapse in reasonable security measures).

The legal framework: DTSA and UTSA

Trade secret protection in the United States runs on two parallel, largely consistent tracks:

• State law — the Uniform Trade Secrets Act (UTSA). Drafted by the Uniform Law Commission in 1979 (amended 1985) as a model statute, the UTSA has been adopted, with state-specific variations, by 49 states and the District of Columbia; New York is the one holdout still relying on its own common-law trade secret doctrine. The UTSA supplies the two-element definition above and the state-court civil cause of action for misappropriation.
• Federal law — the Defend Trade Secrets Act (DTSA) of 2016. Signed into law as Public Law 114-153 and effective May 11, 2016, the DTSA created, for the first time, a federal private civil cause of action for trade secret misappropriation, alongside the existing criminal provisions of the Economic Espionage Act of 1996. It amended 18 U.S.C. Chapter 90 and did not preempt state law — a plaintiff can typically bring both a DTSA claim and a state UTSA claim over the same facts. Its statutory definition, at 18 U.S.C. § 1839(3), tracks the UTSA’s two elements closely (secrecy plus reasonable protective measures, plus independent economic value from the secrecy) and is the definition most U.S. research institutions and TTOs now cite directly.

Both frameworks define “improper means” of acquiring a trade secret — theft, bribery, breach of a duty of confidentiality, electronic espionage — and both explicitly exclude reverse engineering and independent derivation from that definition. This exclusion is the central practical limit of trade secret protection: if a competitor lawfully reverse-engineers a product or independently arrives at the same process, there is no misappropriation and no remedy, no matter how carefully the original secret was guarded.

Trade secret vs. patent: a genuinely different trade-off

A trade secret and a patent are not two versions of the same protection — they trade opposite things for opposite benefits, and the choice between them is a real decision a technology transfer office (TTO) has to make deliberately for each invention, not a formality.

	Trade secret	Patent
Disclosure	None — the entire value depends on the information staying confidential.	Full public disclosure of how to make and use the invention, in exchange for the exclusive right (the patent specification must enable a person skilled in the art to practice it).
Duration	Indefinite — lasts as long as secrecy and economic value are maintained; can in principle last forever (the classic example cited across the IP literature is a closely guarded beverage formula that has never been patented).	Fixed term — 20 years from the earliest U.S. non-provisional filing date for a utility patent, under 35 U.S.C. § 154(a)(2), after which the invention enters the public domain regardless of what the owner wants. See CASRAI’s guide on how long patents last for the maintenance-fee, term-adjustment, and term-extension mechanics that can shift that baseline.
Registration / cost	No filing, examination, or maintenance fees — protection exists automatically once both statutory elements are met; the ongoing cost is the security program itself (access controls, confidentiality agreements, employee exit procedures), not a government fee.	Filing, examination, issue, and periodic maintenance fees; typically one to several years of patent prosecution before a right issues at all.
What defeats it	Independent discovery or lawful reverse engineering by anyone — no infringement claim exists against a party who gets there independently.	Nothing but expiration, invalidity, or non-payment of maintenance fees — a patent blocks even someone who invents the same thing independently and later.
Novelty requirement	None — a trade secret does not need to be new in the patent sense, only secret and valuable. (Contrast with the statutory novelty bar a patent claim must clear under 35 U.S.C. § 102.)	Must be novel and non-obvious over the prior art as of the effective filing date.

Why a TTO might choose trade-secret protection over patenting

Patenting is the default instinct in most university technology transfer offices, but it is not automatically the right call for every invention disclosure. Trade secret protection is generally the better fit when an invention has a characteristic patenting doesn’t reward well:

• The invention can’t be reverse-engineered from the product itself. A manufacturing process, an internal formulation, a proprietary dataset, or a tuning/calibration method that never leaves the lab or the factory floor is exactly the profile where secrecy works and disclosure doesn’t help the inventor. A patent claim on the same process would hand a competitor a roadmap; a trade secret does not.
• The subject matter may not be patent-eligible, or patenting it is a poor return on cost. Software algorithms, internal know-how, negative results (“we tried X and it doesn’t work, don’t waste time on it”), and process optimizations often fail patentability’s novelty/non-obviousness bar, or would cost more to prosecute and enforce than the resulting exclusivity is worth.
• The invention has a short useful commercial life. If a process or formulation will be technologically obsolete well before a 20-year patent term would run out, or before prosecution even concludes, the disclosure cost of patenting buys little.
• The institution wants to avoid starting the disclosure clock. Filing a patent application publishes the invention (typically 18 months after the earliest filing) and puts a hard deadline on the decision to pursue protection at all — see CASRAI’s guide on novelty and invention-disclosure timing under 35 U.S.C. § 102 for how public disclosure can also destroy an inventor’s own ability to patent later. Trade secret protection defers that choice indefinitely.

In practice, many TTOs run a mixed strategy on a single technology: the patentable, disclosure-tolerant parts of an invention go into a patent application, while the surrounding process know-how, manufacturing tolerances, or software implementation details that make the invention actually reproducible stay protected as trade secrets and are transferred to a licensee only under contract. This is a recognized, common pattern in university-industry licensing — see CASRAI’s guide on patent licensing for how exclusive licenses are frequently structured to carry both a licensed patent and an accompanying trade-secret know-how grant.

The role of NDAs in protecting a trade secret

Because trade secret status depends entirely on “reasonable efforts to maintain secrecy,” the confidentiality agreement (NDA) is not incidental paperwork — it is one of the core legal mechanisms that keeps the protection alive. Courts assessing whether an owner took reasonable measures routinely look at whether the information was shared only under a signed confidentiality obligation, in addition to access controls, marking practices, and employee training. A TTO that discloses unpatented process know-how to a prospective licensee, sponsor, or collaborator without an NDA in place risks two distinct harms: the immediate risk that the recipient uses or discloses the information without restriction, and the structural risk that the disclosure itself later gets used to argue the information was never really kept secret, undermining the trade secret claim entirely.

This is why NDAs are routinely executed before, not after, any substantive technical discussion in a licensing negotiation, sponsored-research scoping conversation, or material transfer — see CASRAI’s dictionary entry on the confidentiality agreement (NDA) for the standard terms (permitted purpose, duration, return-or-destruction obligations) those agreements carry in a research-administration context.

Worked examples

• A university materials-science lab develops a novel catalyst. The catalyst’s chemical composition is disclosed in the resulting patent application and issued patent. The precise reaction temperature profile, mixing sequence, and purification steps that make the catalyst reproducible at scale — details the university judges would be extremely difficult for a competitor to reverse-engineer from the product alone — are withheld from the patent specification’s illustrative examples where legally permissible and instead licensed to the manufacturing partner only under an NDA and a separate know-how license.
• A campus software group builds an internal data-cleaning pipeline with tuning parameters refined over years of use on institutional datasets. The parameters are not disclosed in any publication or patent; the group treats the tuned configuration as a trade secret, restricts access to the team, and requires an NDA before sharing it with an external evaluator, rather than filing a patent that would have to publish the details to claim them.

Counter-example

A research team publishes its full method, including all process parameters, in a peer-reviewed journal article before any licensing discussion takes place. Once published, the information is generally known and readily ascertainable — it no longer meets the secrecy element of the DTSA/UTSA definition, and no NDA executed afterward can restore trade-secret status to information the owner has already made public.

Frequently asked questions

Can the same invention be protected by both a patent and a trade secret?

Not the same information at the same time — a patent requires disclosure of what it claims, which is the opposite of secrecy. But it is common and legitimate for one invention to have patentable and non-patentable (or intentionally unpatented) components, with the patent covering the disclosed, claimed portion and a trade secret covering the surrounding know-how needed to actually practice it at commercial scale.

How long does trade secret protection last?

Indefinitely, in principle — there is no statutory term. Protection ends only when the information stops meeting the DTSA/UTSA definition: it becomes generally known, is independently developed by someone else, is lawfully reverse-engineered, or the owner stops taking reasonable measures to keep it secret.

Does an NDA by itself create a trade secret?

No. An NDA is evidence of one of the two required elements (reasonable efforts to maintain secrecy) but does not, by itself, establish that the underlying information has independent economic value from being secret. Both elements of the DTSA/UTSA definition must be met independently of whether an NDA exists.

Is reverse engineering a trade secret misappropriation?

No. Both the DTSA and the UTSA expressly exclude reverse engineering and independent derivation from the definition of improper means. This is the key practical limitation of trade secret protection compared to a patent, which blocks even independent reinvention.

Related terms and guides

See also CASRAI’s dictionary entries on the confidentiality agreement (NDA) and patent prosecution, and its guides on how long patents last, patent licensing, provisional patent applications, patent novelty and invention-disclosure timing under 35 U.S.C. § 102, open-source software licensing in technology transfer, and industry-university research partnership agreement structures, plus the Technology Transfer & Innovation cluster hub.