{/* Schema recommendation: BlogPosting + FAQPage. Set author to Nalin Vahil, datePublished and dateModified to 2026-06-03. The "Frequently asked questions" section maps to FAQPage entities (also provided in the faq frontmatter). The statutory-grounds table and the 5-point self-check table are supporting structured content for the BlogPosting. */}
By Nalin Vahil | Last updated: June 3, 2026 | Cada Grant Strategy
If you've started googling "march-in rights SBIR" after seeing headlines about the government going after Harvard's patents, here's the short version. March-in rights let a federal agency force a patent holder to license a federally funded invention to someone else. They exist under the Bayh-Dole Act. In roughly 45 years, no agency has ever exercised them. The 2025 Harvard proceeding is about alleged compliance failures at a large institution, not the government coming for routine startup patents.
That's the bottom line. The rest of this explains why, what actually triggers march-in, and the five compliance checks that decide whether your startup has anything to worry about.
The law firms publishing alarming client alerts right now write for universities and pharmaceutical companies. Almost nobody is writing for founders. So let's fix that.
What are march-in rights?
March-in rights are a provision of the Bayh-Dole Act (codified at 35 USC 203) that lets the federal agency which funded an invention require the patent owner to grant a license to a third party. If the owner refuses, the agency can grant the license itself. Crucially, march-in forces a license. It does not seize or cancel your patent.
You keep ownership. What you could lose, in the worst case, is exclusivity, because someone else gets a license to practice the invention.
That distinction matters. "The government can take my patent" is the fear. "In a narrow set of circumstances that have never once been triggered, the government could require me to license it" is the reality.
Why is everyone suddenly worried about march-in rights?
Because of one letter. On August 8, 2025, the U.S. Department of Commerce, under Secretary Howard Lutnick, sent Harvard a public letter announcing a comprehensive review of Harvard's federally funded research and asserting a march-in proceeding against the university's patents (source: Wilson Sonsini, Duane Morris client alerts).
The letter gave Harvard until September 5, 2025 to hand over a list of every patent arising from federal funding, plus proof it had complied with Bayh-Dole.
This is unprecedented. In the roughly 45 years since Bayh-Dole passed in 1980, no federal agency had ever pushed a march-in proceeding to this stage. That's why the legal world lit up. And that's why founders who happen to be applying for an SBIR right now are nervous.
Here's the problem with the coverage: it's aimed at general counsels of $50 billion endowments, not at a seed-stage founder with one provisional patent and 6 months of runway. The advice does not translate.
What the Harvard case is actually about
This is the part the client alerts bury. The Commerce Department did not say "Harvard built something valuable, so we want it." It alleged three specific compliance failures (source: Wilson Sonsini analysis of the August 8, 2025 letter):
- Late invention disclosure and election of title -- not reporting inventions to the funding agency on time (35 USC 202(c)(1)).
- U.S. manufacturing preference -- failing to ensure inventions licensed exclusively for the U.S. market are substantially manufactured in the U.S. (35 USC 204).
- Failure to achieve practical application -- not taking effective steps to actually commercialize the funded inventions (35 USC 203(a)(1)).
Read that list again. Every item is a process and compliance allegation. None of them is "you have a patent the government wants."
Think about the difference in concrete terms. Imagine a fictional university tech-transfer office sitting on hundreds of federally funded patents, some disclosed late, some licensed to overseas manufacturers, some never developed. Now imagine a fictional 8-person robotics startup that disclosed its one invention to the agency 6 weeks after filing, plans to build in Ohio, and is actively shipping product. Those two are not exposed to march-in in remotely the same way.
The Harvard case is not a signal that the government is hunting startup IP. It's a signal that compliance basics, which were treated as paperwork for 45 years, are now being read closely.
Have march-in rights ever actually been used?
No. Not once. In the roughly 45 years since Bayh-Dole became law, no federal agency has ever exercised march-in rights (source: Congressional Research Service; GAO).
The NIH alone has received and denied every march-in petition it has gotten. The most cited example: in 2004, petitioners asked NIH to march in on the HIV drug ritonavir (Norvir) over its price. NIH said no, because the drug was available to the public. Similar petitions on other drugs were also denied.
In about February 2026, the Government Accountability Office added hard numbers to the debate. The GAO found that using march-in to lower drug prices would have minimal practical impact: as few as 18 of the 883 FDA-approved drugs since 1985, roughly 2%, are covered entirely by government-interest patents that could even qualify (source: GAO report, via Jones Day and Mondaq summaries). The GAO also warned that price-based march-in could create a "chilling effect" on commercializing federally funded inventions.
Now the honest caveat. "Never been used" was a clean talking point for 45 years. The Harvard proceeding makes it "never been used, but one is now in motion." Frankly, nobody knows yet how the Harvard matter resolves. What we do know is what the law actually requires, and that's what you can control.
What actually triggers march-in rights?
Government march-in rights are not open-ended. The statute (35 USC 203) lists four grounds, and an agency can march in only if one of these applies:
| Statutory ground | What it means in plain English | How likely for a startup? |
|---|---|---|
| Practical application not achieved | You sat on the invention and never developed or commercialized it | Low if you're actively building |
| Health or safety needs unmet | A public health/safety need isn't being met and the invention could meet it | Low for most tech; relevant mainly to drugs/devices |
| Federal public-use requirements unmet | Specific federal use regulations aren't being satisfied | Rare; program-specific |
| U.S. manufacturing preference breached | You gave an exclusive U.S. license to a foreign manufacturer without a waiver | Controllable -- a contracting decision |
Notice what's not on the list: "your patent is worth a lot," "you raised a big round," or "you won an SBIR." None of those is a trigger. Every ground is about non-performance or non-compliance, not success.
The real risk for your startup is compliance, not confiscation
Here's the reframe that the legal fear-mongering misses. Your march-in exposure is almost entirely a function of compliance hygiene, and compliance hygiene is controllable.
There are five Bayh-Dole obligations that, kept clean, push your march-in risk close to zero. These are the same five we built into a risk checklist for Cada's roadmap process.
- Timely invention disclosure (35 USC 202(c)(1)). Disclose each subject invention to the funding agency within 2 months of your inventor disclosing it internally. Missing this window is the single most common startup mistake.
- Election of title and filing. Elect to keep title within 2 years of disclosure, and file your patent application within 1 year of electing. Calendar these dates the day funding lands.
- U.S. manufacturing preference (35 USC 204). If you grant an exclusive license to sell the product in the U.S., the product generally has to be substantially manufactured in the U.S. (or you request a waiver). This is a contracting decision you make with eyes open.
- Utilization reporting (35 USC 202(c)(5)). File the periodic utilization reports the agency asks for. Ignoring them is a self-inflicted flag.
- Government license and funding statements (35 USC 202(c)(4)). Acknowledge the government's nonexclusive, paid-up license, and put the correct federal funding statement in your patents.
None of this requires a law firm on retainer. It requires a calendar, a spreadsheet, and the discipline to treat compliance deadlines like fundraising deadlines.
A 5-point march-in risk self-check
Run your own company through this. Score each line low, medium, or high. We're giving away the exact checklist we use internally, because the diagnosis is free and the doing is where founders ask for help.
| # | Compliance dimension | Ask yourself | Risk if "no" |
|---|---|---|---|
| 1 | Timely disclosure | Did we report each funded invention to the agency within 2 months? | High |
| 2 | U.S. manufacture | Are we keeping U.S. manufacturing for any exclusive U.S. license (or did we get a waiver)? | Medium-High |
| 3 | Practical application | Are we actively developing and commercializing, not shelving the invention? | Medium |
| 4 | Utilization reporting | Are we filing the agency's periodic utilization reports? | Medium |
| 5 | Government license | Do our patents carry the correct federal funding statement and acknowledge the government license? | Low-Medium |
If every answer is a confident "yes," your march-in risk is about as low as it gets, and the Harvard headlines are not about you. If you've got two or three "no" answers, that's not a crisis, but it's a reason to clean up your compliance before you file your next application.
What to do if you're worried about march-in rights
Concrete steps, in order:
- Calendar your three deadlines now: 2-month disclosure, 2-year election, 1-year filing. Late paperwork is the only Harvard allegation that's purely on you to prevent.
- Document your U.S. manufacturing intent. If you plan to manufacture abroad under an exclusive U.S. license, talk to counsel about a waiver before you sign anything.
- Keep utilization records. A simple log of how the invention is being developed answers the "practical application" question before it's ever asked.
- Don't let a 2% statistic and a never-once-used law scare you off non-dilutive funding worth up to $314K (and more at some agencies) in Phase I.
And the honest part: the Bayh-Dole enforcement posture genuinely shifted in 2025, and edge cases (foreign manufacturing, exclusive licensing structures, dormant patents) deserve a real conversation with IP counsel. For the typical seed-stage SBIR applicant who discloses on time and builds in the U.S., the answer is reassuring. For the edge cases, get specific advice.
Frequently asked questions
Can the government take my patent if I win an SBIR?
No. March-in forces a license to a third party in narrow circumstances. It never transfers ownership, and it has never been exercised in roughly 45 years of Bayh-Dole.
Does winning an SBIR put my IP at higher march-in risk than other companies?
No. The same Bayh-Dole rules apply to every recipient of federal research funding. An SBIR award doesn't single you out. It just means the standard compliance obligations apply.
Should the Harvard case change whether I apply for SBIR?
No. It should make you take the compliance basics seriously, especially the 2-month invention disclosure window. It is not a reason to skip up to $314K (and more at some agencies) of non-dilutive Phase I funding.
What's the single biggest march-in compliance mistake startups make?
Missing the invention disclosure deadline. You have 2 months from your inventor's internal disclosure to report a subject invention to the funding agency. Blow that, and you've created exactly the kind of failure cited against Harvard.
Is pricing my product "too high" a march-in trigger?
Contested, and never successful. Petitioners have repeatedly asked agencies to march in over drug prices, and every petition has been denied. The GAO's 2026 report questions whether price-based march-in would even matter, finding only about 2% of approved drugs could qualify.
Get a straight answer on your Bayh-Dole compliance
If you're applying for an SBIR and you're not sure whether your IP compliance is clean, that's a 15-minute conversation, not a 6-month worry.
Cada has helped startups structure their grant compliance so their IP is protected from day one, across 100+ proposals and 30+ agencies. We'll run your company through the same 5-point march-in checklist above and tell you straight whether you have anything to fix. No pitch, no obligation.
Want the full framework first? Read our guide on SBIR IP Protection: What the Bayh-Dole Act Actually Means for Your Startup for the complete picture on how Bayh-Dole handles your patents.