Neurotech has a vaporware problem for a simple reason: the real thing is spectacular. 🧠 Paralyzed people move cursors with their thoughts, and that is a matter of record. So a startup with a sleek render and a founder who talks fast can borrow that credibility without owning any of the proof.
This is a checklist for readers, investors and journalists. Every red flag below is a signal, not a verdict. A company can trip one for innocent reasons, and a real one can be early, quiet and awkward on stage. I use named companies only as reference points for what the paper trail looks like when it exists, and where I cite an enforcement action, I report what the agency alleged and nothing more. The test throughout is the one this publication cares about: does the story separate a lab demo, an early clinical trial, a cleared product, and something years away? How many pitches you have seen this year do that on slide one?
Red flag 1: a designation gets sold as an approval
Regulatory vocabulary is a minefield, and some founders wade in on purpose. The one to know is Breakthrough Device designation. The FDA states that it is neither an approval nor a clearance. It is a program status, granted before a marketing submission, that buys prioritized review and more frequent talks with the agency. The device still has to clear the same safety and effectiveness bar. 🧾
Real companies hold designations and say so carefully. Neuralink’s implant has Breakthrough designation for speech restoration and its Blindsight project, and it has 26 patients as of June 2026, in trials across the US, UAE, UK and Canada. None of that is a marketing authorization. 🚧 Synchron holds a designation from August 2020 and, per The Robotics Media, is preparing the first trial meant to support a PMA for a permanently implanted BCI.
Green flag: the deck names the pathway (IDE, De Novo, 510(k), PMA) and the stage in the same sentence
Yellow flag: “FDA-backed,” “FDA-recognized” or “FDA-fast-tracked” with no pathway named
Red flag: a designation, a meeting or a registration presented as if it were authorization
Even real clearances have limits. Precision Neuroscience announced 510(k) clearance in April 2025 for its cortical electrode array, for recording, monitoring and stimulation of brain-surface activity, with implants of up to 30 days. That is a real authorization for a temporary, removable device. ✅ It is not a permanent home BCI, and a headline that blurs those two is doing the blurring on purpose.
Red flag 2: no trial anyone can look up
Real neurotech leaves a paper trail, because regulators, ethics boards and hospitals demand one. A company with human data should be able to tell you the design, the number of participants, the endpoints, and where to find the registry entry on ClinicalTrials.gov. If the evidence is a video of one impressive session, it is a demo. Demos are fine. Calling them a trial is the problem. 🎥
Look at what a modest, honest result looks like. Synchron’s COMMAND early feasibility study enrolled six patients and, as of September 30, 2024, all six met the primary safety endpoint over 12 months with no serious adverse events involving the brain or vasculature. Six people is small. 🔍 It is also specific, checkable and about safety, which is what an early study is supposed to be about.
Ask for: sample size, control condition, primary endpoint, and follow-up length
Watch for: “thousands of users” standing in for a trial, or self-reported surveys described as a study
Discount: results that appear only in the company’s own press release, until a peer-reviewed paper exists
Funding is not evidence either, but it is context worth having. Synchron announced a $200 million Series D in November 2025, and that money follows a trial record instead of replacing one. If you want to see who is funding what, and how a startup’s papers, rounds and trials line up, NeurotechMag Pro members can track every paper, funding round, and trial connected to developments like these, summarized and searchable. 📈
Red flag 3: “wellness” on the box, medicine in the marketing
In the US, a company can sell an EEG headband, a focus tracker or a sleep earbud without FDA review if it stays inside the general wellness lane, which NeurotechMag’s myths piece explains well. The January 2026 guidance, as summarized by the National Law Review, keeps claims about stress, sleep and fitness inside the lane. Claims to cure, mitigate, prevent or treat a disease, “medical grade” language, and prompts to change treatment fall outside it. ⚖️
The red flag is a mismatch between two things, and regulators read the whole page: a legal disclaimer that says one thing and a landing page that says another.
The FTC’s Lumosity case is the classic precedent. In January 2016 the company agreed to pay $2 million, after the FTC alleged it claimed its brain games could delay age-related decline and reduce impairment from conditions such as stroke and ADHD. The FTC’s line: “Lumosity simply did not have the science to back up its ads.”
In 2022 the FDA sent a warning letter to NeuroField, alleging it marketed neurostimulation and EEG devices without the required premarket approval, clearance or investigational exemption. The company said its devices were biofeedback tools for relaxation training and muscle reeducation, and that it was working with the FDA. Those are allegations and a response, not a finding.
A good sign is boring language: what the product measures, what it does not do, and no disease words
Neither case proves anything about a startup you are evaluating today. 🕵️ They show how the lane is policed, which is enough to tell you what a careful company avoids saying. 🚦
Red flag 4: the specs are missing, or only the flattering one is shown
This audience wants numbers, and honest companies publish them. Electrode or channel count, signal bandwidth, latency, invasive or not, lead investigator, and stage. A pitch that says “advanced AI decodes your brain” and skips all six is hiding something, or has nothing to hide and nothing to show.
Beware the single headline number as well, because bigger is not automatically better. ⚠️ Synchron’s Stentrode has 16 electrodes and goes in through a blood vessel, which The Robotics Media calls a deliberate trade-off: lower signal fidelity in exchange for a procedure built on routine interventional techniques. A company that explains its trade-offs is one worth taking seriously. 🔬
On the consumer side, specs are a useful honesty test. NeurotechMag’s guide to devices you already use describes Neurable’s MW75 Neuro headphones with 12 dry EEG channels and a listed bandwidth of 0 to 131 Hz. A stated bandwidth can be checked, argued over and tested by someone else. “Proprietary brain score” cannot.
Ask: was the metric validated against a clinical-grade reference, and by whom?
Ask: can you export raw data, or is everything locked in an app?
Ask: is the “focus” or “calm” score described in a peer-reviewed paper?
If the answer to all three is “trust us,” that is your answer. 💡 Would you buy a car on those terms? 🚗
Red flag 5: the privacy policy says more than the science does
Brain data is personal in a way most consumer data is not, and its value to a startup is obvious. So check what the company does with the signal before the science is proven. 🔒
The Neurorights Foundation reviewed consumer neurotech companies and found that, per KFF Health News, 29 of 30 with online products have access to brain data, and that they place no meaningful limits on that access. 🔒 Nearly all can share it with third parties. Colorado passed a neural data law in 2024, California has followed, and Montana’s took effect in October 2025.
Green flag: a plain-language policy on who can access raw and derived brain data, and how deletion works
Yellow flag: neural data lumped in with generic “usage data” and shared with unnamed partners
Red flag: a product that needs your brain signal to work and a policy that will not say where it goes
NeurotechMag’s seven-question buying guide covers this too, along with the boring but decisive matters of full 12-month cost and whether the company will still exist in a year. A startup can be sincere and still fail one of these tests. That is fine, so long as you know which one. 🎯
So use the five as a filter, not a courtroom. Score each pitch: pathway named, trial findable, claims match the lane, specs published, data terms clear. Which of the five does the next startup in your inbox fail first?