On February 12, 2026, the FTC's noncompete ban stopped existing even on paper. That morning, the Federal Trade Commission (FTC) published a final rule deleting its own noncompete rule from the Code of Federal Regulations, tidying the books to match a court decision that had already been reality for eighteen months.
The FTC's own summary now reads like an epitaph: "The Noncompete Rule is not in effect and it is not enforceable."
The ban that never was
- April 23, 2024. The FTC issues a final rule banning nearly all new noncompetes and voiding most existing ones, effective September 4, 2024.
- August 20, 2024. Two weeks before the effective date, the U.S. District Court for the Northern District of Texas rules in Ryan, LLC v. FTC that the agency lacked the statutory authority to issue the rule at all, and sets it aside nationwide.
- September 5, 2025. Under new leadership, the FTC votes 3-1 to drop its appeal and accept the ruling.
- February 12, 2026. The rule is formally removed from the Code of Federal Regulations.
A nationwide ban could theoretically return through a brand new rulemaking or an act of Congress. The vehicle that was supposed to deliver one, the Workforce Mobility Act, was reintroduced in the Senate in June 2025 and has been sitting in committee since. Don't hold your breath.
What the ban was supposed to fix
The diagnosis was never the controversial part. An estimated 30 million people, nearly one in five American workers, are subject to a noncompete. When the FTC asked for public comment on the proposed rule, more than 26,000 comments came in and over 25,000 supported the ban. Rulemakings don't usually get that kind of consensus.
The FTC projected the rule would raise earnings for the average covered worker by $524 a year, grow new business formation by 2.7% annually, and produce more than 8,500 additional startups every year. The logic: noncompetes trap people in jobs they've outgrown, suppress wages, and scare would-be founders out of starting companies in the industries they know best.
And the practice had drifted a long way from its original purpose. The concept was built for executives with real trade secrets. By the mid-2010s it was being applied by reflex to people making sandwiches. Jimmy John's had its sandwich makers and delivery drivers signing agreements that barred them, for two years after leaving, from working at any establishment within two miles of a Jimmy John's location that made more than 10% of its revenue from sandwiches. The New York attorney general declared those agreements unlawful in 2016, and the franchisees that had used them agreed to void them.
A tool designed for the C-suite had become a way to keep a sandwich artist from defecting to the sub shop across the street.
The agreement that does the actual work
The most useful thing the FTC produced in 2024 sits some eighty pages past the ban that got all the attention, in Part IV.D.2 of the final rule, under the heading "Employers Have Alternatives to Non-Competes for Protecting Valuable Investments." It is the section where the agency explains why employers never needed the clause it was banning, and the argument is short enough to quote.
Trade secret law, the Commission found, "provides employers with a viable, well-established means of protecting investments in trade secrets, without the need to resort to the use of non-competes with their attendant harms to competition." Non-disclosure agreements (NDAs) are "another well-established, viable means." And the comparison between the two tools is the part that should interest anyone still writing noncompetes into a template:
Appropriately tailored NDAs burden competition to a lesser degree than non-competes. Such NDAs may prevent workers from disclosing or using certain information, but they generally do not prevent workers from seeking or accepting other work, or starting their own business, after their employment ends.
Most employers already have both. In the survey the FTC leaned on, 95.6% of workers with a noncompete also had an NDA, and by the Commission's own arithmetic on that data, 97.5% had an NDA, a non-solicitation agreement, or a non-recruitment agreement. (The survey is a self-selected sample of Payscale visitors rather than a census, a caveat the FTC itself flagged.) The noncompete is the only one of those instruments that fires when nothing has been taken.
Employers push back on this, and the sharpest version of the objection came from the U.S. Chamber of Commerce, in its 2023 comment to the FTC. The confidential information a business wants to protect is broader than trade secret law reaches, misappropriation is hard to prove, and, quoting testimony from the FTC's own forum, a departing worker "cannot excise [a company's] confidential information from her brain." The Chamber also pointed at the bill, citing a 2019 survey that put the median cost of litigating a trade secret case at $4.1 million when $10 million to $25 million is at stake. The clock is worse than the invoice: Lex Machina counts a median of 1,124 days from filing to trial for federal trade secret cases tried between 2023 and 2025, a long wait for a remedy to a harm that landed on day one.
The FTC's answer:
To the extent trade secret law and NDAs require higher evidentiary showings, that makes these alternatives more tailored tools for protecting employers' valuable investments without unduly restricting a worker from engaging in competitive activity.
The higher bar is not a bug. It is the filter. A noncompete requires you to prove only that someone went to work for a competitor. An NDA or a trade secret claim requires you to prove that something of yours actually walked out the door. The first standard sweeps in every employee who simply got a better offer; the second catches the case you were actually afraid of. And the head start you give up is smaller than it looks, because the Commission noted that prophylactic relief is already available in the nearly half of US states that recognize the inevitable disclosure doctrine, which lets an employer move before the secret has demonstrably walked.
There is now field evidence that the difference is real. In a randomized experiment at a single firm, forthcoming research found that adding a noncompete to a job offer cut the odds of a worker moving to the named competitor by 57%, and had no effect at all on whether they shared information covered by their NDA. The clause changed where people went; it did not change what they disclosed. Separately, a study of state noncompete bans found that when mobility went up, trade secret litigation stayed flat in the short run and fell in the long run. Freeing people to move did not produce the theft wave the clause is supposed to prevent.
Congress agrees more than most employers realize. The Defend Trade Secrets Act of 2016 gave trade secret owners a federal cause of action with injunctions, up to double damages for willful misappropriation, and fees. It also wrote in an explicit limit: a federal injunction may not "prevent a person from entering into an employment relationship," and any conditions on that employment must rest "on evidence of threatened misappropriation and not merely on the information the person knows." The country's strongest trade secret statute deliberately refuses to do the one thing a noncompete does.
Where this argument goes wrong
If NDAs are the fairer tool, that is only true of NDAs drafted to be fair, and a lot of them are not.
The FTC said so in the same rule. An NDA can be written so broadly that it "functions to prevent a worker from seeking or accepting employment or operating a business," at which point it is a noncompete wearing a different hat. The examples the Commission gave: an NDA barring disclosure of any information that "relates to" the industry, or one covering everything the worker learned on the job including publicly available facts. The First Circuit has called those agreements out for raising "the same policy concerns about restraining competition as noncompete clauses." A 2024 Yale Law Journal study of confidentiality agreements produced in trade secret litigation documents agreements that cover generally known information, prohibit use as well as disclosure, and, unlike most noncompetes, carry no geographic or time limit at all. Noncompetes get judicial scrutiny in every state; overbroad NDAs mostly don't, which is exactly why an employer tempted to replace one with the other should resist writing the second one like the first.
The line the FTC drew is a good drafting test, and you can apply it without a law degree. An NDA stays an NDA when its restrictions do not reach information that "arises from the worker's general training, knowledge, skill or experience," or that is "readily ascertainable to other employers or the general public."
States are starting to put teeth behind that line. Colorado voids a confidentiality provision that reaches a worker's general training, knowledge, skill, or experience, or information readily ascertainable to the public, treats it as an illegal restrictive covenant, and exposes the employer to a $5,000 penalty per worker merely for presenting one, though courts can waive the penalty for a good-faith mistake. California goes further, and this is the exposure most remote employers are quietly carrying. Since January 1, 2024, under Business and Professions Code sections 16600.1 and 16600.5, an employer that so much as includes a void noncompete in a California employee's contract commits a civil violation and an act of unfair competition, and the employee has a private right of action with attorney's fees. The same legislation required employers to send individual written notice, by February 14, 2024, to every current employee and every former employee employed after January 1, 2022, telling them their noncompete is void. California courts have also struck overbroad confidentiality clauses as de facto noncompetes. If your template still carries the clause and you have California headcount, which for a remote employer is nearly a given, the cleanup is already overdue.
One small, free thing while you are in the document: the Defend Trade Secrets Act requires a whistleblower immunity notice in any agreement governing trade secrets or confidential information. Leave it out and you forfeit the double-damages remedy and attorney's fees against that worker. It costs a paragraph.
What governs now: the map
With the federal rule gone, the map is the law, and most of it already points one direction. Four states, California, Minnesota, North Dakota, and Oklahoma, void nearly all employee noncompetes, and in three of them the prohibition dates to the 1800s; a broad ban of Wyoming's own took effect in July 2025, prospective and with an executive carve-out. Thirteen states plus D.C. void the clause below employee pay thresholds running from roughly $30,160 in New Hampshire to $162,164 in D.C. Healthcare is moving fastest of all: 48 of the 110 noncompete bills introduced across 34 states in 2026 were healthcare bills. Florida went the other way, with a CHOICE Act that allows noncompetes up to four years and makes preliminary injunctions nearly automatic. Florida is the outlier.
The change that should actually alter behavior has not landed yet. Washington passed the broadest ban of 2026: signed in March and effective June 30, 2027, it will void nearly all noncompetes for employees and independent contractors, including agreements already signed. Damages will be the greater of actual harm or $5,000 plus fees, and simply telling a worker they are bound by a void clause will itself be a violation. If you hire in Washington, that is a date for the calendar rather than a headline to file away.
For a distributed employer, the map raises a question none of the state trackers answer: which state's law applies to which person? The agreement signed in Texas by someone who now lives in San Diego, the Delaware forum-selection clause attached to a Washington hire, the offer letter that predates a relocation. California declares a noncompete that is void under its law unenforceable "regardless of where and when the contract was signed," and Washington already voids, for Washington-based workers, any clause requiring the covenant to be adjudicated out of state or under another state's law. Where your people are today governs more than where they were when they signed, and for a remote workforce that answer changes every quarter. It is a question for counsel, not a blog post.
Outside the United States the pattern is different but the lesson is the same. There is no European Union ban, and no proposed one; the EU's 2016 Trade Secrets Directive goes out of its way to say that "Nothing in this Directive shall be understood to offer any ground for restricting the mobility of employees." What much of continental Europe did instead was put a price on the clause, with Germany the canonical example: at least 50% of the worker's prior compensation for every year of the restriction, capped at two years. In July 2026, the OECD published the first systematic cross-country evidence on noncompetes, covering fifteen countries. Pricing the clause did not reduce its use: prevalence runs between a fifth and a third of the workforce, and "the prevalence of non-compete clauses does not vary systematically across regulatory regimes." In the countries where compensation is legally required, almost half of the firms using noncompetes pay nothing to any of the workers bound by them, and 31% of firms that use noncompetes apply them to every employee regardless of role.
Nobody, anywhere, has landed on a clean answer. And losing the rule did not end the FTC's interest: in place of a blanket ban, the agency now pursues case-by-case enforcement under Section 5 of the FTC Act. In September 2025 it moved against Gateway Services, the country's largest pet cremation company, over noncompetes binding nearly 1,800 workers, finalizing the order that November, and days after the first move sent warning letters to large healthcare employers and staffing firms. In April 2026 it went after Rollins, the pest-control giant behind Orkin, producing a final order in June 2026 that frees more than 18,000 workers, plus warning letters to thirteen other pest-control companies. The blanket rule is gone, but an aggressive noncompete can still earn you a letter from Washington.
What the clause costs before anyone enforces it
The legal exposure is counsel's problem. The quieter cost is yours, and it starts before any of this reaches a courtroom. Candidates read a noncompete as a signal about how you think about people leaving; some negotiate around it, some price it into the offer, and a few walk. Recruiters get asked "is this enforceable?" and have no good answer. Alumni you would happily rehire hesitate to come back, and a referral network is harder to build when leaving carries a tail. None of that appears as a line item, which is how the clause survives budget review while quietly taxing offer acceptance. And if you do drop it, say so out loud, in the job post and the offer call. In a market where the default is suspicion, "we don't do noncompetes" costs nothing to say and is one of the few trust signals a candidate can check.
What to do with this
If you're a People leader looking at your template agreements in 2026, treat this as prep for a working session with counsel, with the ownership split explicit: the messaging above is entirely yours; the inventory and questions below are yours to start and counsel's to finish. Four pieces of homework:
Two things to do this week. First, open your NDA and confidentiality templates and check for the DTSA whistleblower immunity notice; if it's missing, adding it is a paragraph of standard language and it restores remedies you have already paid for. Second, if your template has ever included a noncompete and you have California headcount, confirm the February 2024 notice obligation was actually met, because that liability is already running, not hypothetical.
The inventory, with the actual columns. One owner, usually People ops with counsel on call, a timeline in weeks, and a spreadsheet with a row per person and these columns: role; state (or country) at signing; state today; every restrictive covenant they signed, meaning noncompete, NDA, customer non-solicit, employee non-recruit, and any forfeiture-for-competition terms in the equity documents; signing date; the contract's governing-law and forum clause; and compensation against the local threshold. Most companies cannot produce this without digging, which is the point of producing it.
Match the instrument to the risk, then test the instrument. For each role, name what you would actually be afraid of losing, and reach for the narrow tool built for it: customer relationships take a customer non-solicit, team raids take a non-recruit, secrets take the NDA plus trade secret law, and for the handful of executives who genuinely carry the roadmap, garden leave or equity forfeiture terms do the work noncompetes pretended to. Then check the NDA against the FTC's test above: if it reaches general skill and experience, or publicly known information, or has no time limit, it fails, and in a growing number of states it fails as a matter of law.
Five questions to bring to counsel. Which of our existing noncompetes are already void where the signer lives today, and do any require affirmative notice or rescission? Do our governing-law and forum clauses survive an employee's move to California today or Washington after June 2027? Do our NDAs or non-solicits reach anything the FTC's drafting test says they shouldn't? What does our real exposure look like if we keep the clause for another year? And if we dropped the noncompete from the template tomorrow, what protection, specifically, would we lose that the rest of the stack doesn't already provide?
That last question usually has an uncomfortable answer, and the FTC gave it first: employers that wish to retain workers can "compete on the merits for the worker's labor services," by paying more, offering better hours, or otherwise improving the job. The agency was being pointed about it, and it was right.
Our take
None of this is legal advice, and the patchwork above is exactly why that phrase exists. But the strategic direction doesn't need a law degree: artificial barriers between talent and the best employers hurt everyone except the employers people are trying to leave.
That fight is over for now, and it mattered less than it looked. One agency spent three years and 26,000 comments trying to ban the clause and ended up deleting its own rule, while the tool that reliably protects what employers actually need protected was sitting in the same folder all along, the one that never required telling a person where they're allowed to work next.
So the question stops being what you can stop people from doing, and becomes why they'd stay. That's a harder question, and candidates are already answering it for themselves: they read your policies, your reviews, and your hiring zones before they ever talk to you. A claim you make about yourself is marketing, and candidates know it. Independent verification is the one thing on the page you can't manufacture.
That's the gap our Top Remote Culture certification exists to close. No survey theater: an independent pass/fail check of your policies, benefits, reviews, and hiring practices against a published Remote Culture Standard, with a badge candidates can see and a verified profile they can trust. Your rating stays yours, by the way. No company can buy a higher score. If the best talent is free to go anywhere, give them somewhere verified to go.
