The 2026 BCI Race: Who's Actually Ahead, and How Close Is China?
In Q2 2026, Brain-Computer Interface (BCI) finally graduated from "lab spectacle" to "clinical pathway in sight." Neuralink's PRIME study expanded to ~21 global participants in January. Synchron closed a 1B." China's BrainCo raised ~$286M in January 2026 — the largest single BCI round outside the United States in history.
This piece puts the 14 most-watched companies of 2024–2026 on one map and answers the questions I keep being asked:
- Who actually has long-term human data?
- Which of the four electrode philosophies — intracortical, semi-invasive, endovascular, non-invasive — looks strongest in 2026?
- How fast is China catching up, and is it catching up the same way the US is moving?
- Who's most likely to land the first FDA-approved commercial BCI?
All numbers below come from press releases, SEC filings, ClinicalTrials.gov, Stat News, MedTech Dive, IEEE Spectrum, Nature, Yicai Global, SCMP, etc. Anything supported by only one source is flagged "unverified."
Five takeaways for 2026
-
Patient counts crossed the low double digits, and China is closing fast. Neuralink went from 12 participants in September 2025 to ~21 in January 2026. NeuroXess (China) claims 54 cumulative human BCI implants since 2022. If true, that's more than any single US company. If we filter for permanent intracortical implants with multi-month longevity data, Neuralink still leads, with Synchron close behind on the endovascular side.
-
Semi-invasive and endovascular have become the pragmatic middle lane. Synchron's Stentrode (a stent-based electrode delivered via the jugular vein into the cortical vein over motor cortex), Precision Neuroscience's surface Layer 7 (FDA 510(k) cleared March 2025), and China's NeuCyber Beinao-1 (epidural semi-invasive) all skip craniotomy. The bet: signal quality is "good enough," in exchange for materially lower surgical risk and a faster regulatory path.
-
2024–2025 saw a wave of FDA Breakthrough Device Designations and IDE approvals. Neuralink (N1/Telepathy, Blindsight), Synchron (Stentrode), Blackrock (MoveAgain), Paradromics (Connexus, two designations — paralysis + speech restoration), Precision (Layer 7), and CorTec (stroke motor rehab, April 2026 — the first BCI ever designated for that indication). Paradromics and CorTec both performed first-in-human procedures in 2025.
-
Funding has bifurcated. Neuralink raised a 9B. Synchron raised 1B." Precision raised 286M in January 2026, StairMed has raised 30–125M. Tether took a majority stake in Blackrock Neurotech for $200M in April 2024 — the first time crypto balance sheets flowed into medical devices at scale.
-
There will not be a "consumer" BCI in 2026. Every company today still has a medical-device narrative (paralysis, ALS, stroke, blindness). The closest thing to "consumer feel" is BrainCo's non-invasive prosthetic hand, already in mass production in Hangzhou. On the implant side, Synchron is the most plausible candidate to reach "consumer-adjacent" first — their Apple Vision Pro / iPhone integration was announced in May 2025, and the pivotal trial begins in 2026, with earliest possible US commercial launch in 2027–2028.
US camp: penetrating intracortical
Neuralink — the biggest bet on high-bandwidth
Founded 2016 by Elon Musk and others. HQ Fremont/Austin. Closed a 9B, with secondary-market valuations as high as 1.29B.
The core device is the N1 implant — 1,024 channels across 64 flexible threads. The R1 robot handles thread insertion. Product brands:
- Telepathy — motor / cursor control (live with users)
- Blindsight — visual cortex stimulation for restoring low-resolution vision (FDA Breakthrough Designation Sep 2024)
- Telekinesis — trademarked, on the roadmap
Clinical status: PRIME early-feasibility (NCT06429735) is the lead trial, with CONVOY (robotic arm / assistive control), GB-PRIME (UCLH, 7 participants by late 2025), and UAE-PRIME (NCT06992596) added. By January 2026, ~21 participants worldwide — the fastest ramp in the entire BCI sector.
The two best-known recipients:
- Noland Arbaugh (implanted Jan 2024): early thread retraction reduced channel count to ~15% of baseline; software compensation recovered most cursor performance. Still using the device as of August 2025.
- Brad Smith (first ALS recipient): uses Link plus eye tracking for communication.
Risks: Thread retraction was publicly acknowledged in patient one and only stabilized over months. Device longevity at scale is unproven. Critically, Neuralink has not published a peer-reviewed paper on its human data — a stark contrast to BrainGate, Synchron, Precision, and ONWARD, all of which have Nature / NEJM-class publications.
Synchron — the no-craniotomy platform play
Founded 2012 by Thomas Oxley and Nicholas Opie at the University of Melbourne. HQ now Brooklyn. Series D 1B", led by Double Point Ventures. Total raised ~$345M.
Core device is the Stentrode — a 16-channel electrode array on a stent, delivered through the jugular vein into the superior sagittal sinus, lying directly above motor cortex. Fully endovascular, no open-brain surgery, paired with the Synchron Switch BCI system.
Clinical status: The COMMAND US early-feasibility study (NCT05035823) enrolled 6 participants (plus the earlier Australian SWITCH cohort, ~10 permanent implants total). In September 2024, Synchron announced positive results: no device-related serious adverse events at 12 months, the primary safety endpoint met. CEO Tom Oxley has stated the pivotal trial begins in 2026.
Three 2025 partnerships rewrote the company's narrative:
- Apple (May 2025): Synchron will be the first BCI with native support for Apple's BCI-HID profile — direct brain control of iPhone, iPad, and Apple Vision Pro.
- NVIDIA (GTC 2025): Demoed the Stentrode + Holoscan AI + Vision Pro stack end-to-end.
- OpenAI: Reported integration with ChatGPT through Stentrode, no commercial agreement publicly confirmed.
This is a philosophical pivot for the BCI sector: Synchron no longer sells itself as "an ALS device." It's positioning as the platform BCI for XR and agentic AI. The implant procedure is interventional radiology — thousands of US physicians already do stent work, which is a vastly stronger scalability story than Neuralink's neurosurgery dependency.
Risk: 16 channels is a ceiling. Synchron's signal bandwidth is far below Neuralink or Paradromics, which may cap high-degree-of-freedom (high-DOF) applications.
Paradromics — betting on speech restoration
Founded 2015 by Matt Angle in Austin. Cumulative funding ~88.7M to $127M). Most recent round was a Series A extension in February 2025 with NEOM joining as an investor (the Saudi futuristic-city fund — interesting geopolitical signal).
Core device Connexus: a coin-sized, fully implantable module with 421 microwire electrodes that record from individual neurons. Claimed >200 bits/sec information transfer rate (preclinical).
Clinical status:
- May 2025: First-in-human neural recording during an epilepsy surgery at a University of Michigan partner (intraoperative, not a permanent implant).
- November 2025: FDA cleared Connect-One IDE, study planned for Q1 2026, two participants initially, focused on speech restoration. Paradromics calls this the first IDE for a fully implantable BCI in speech restoration.
- Holds two FDA Breakthrough Device Designations (paralysis + speech restoration).
Differentiator: Highest neuron-level fidelity in the commercial cohort after Neuralink. Everyone else chases cursor / robot arm; Paradromics is betting on speech first. Fully implantable (Utah Array still requires a skull-mounted pedestal).
Precision Neuroscience — getting to FDA without going into the brain
Founded 2020/2021 by Benjamin Rapoport (a Neuralink co-founder who left in 2018 over safety concerns) and others. **Series C 500M; total funding ~$180M.
Core device Layer 7 Cortical Interface: a thin-film electrode array that sits on the surface of the cortex (subdural), not penetrating brain tissue. 1,024 channels per array, inserted through a sub-millimeter cranial slit and conforming to cortical curvature.
March 2025: Received FDA 510(k) clearance (K242618) — the first high-channel-count BCI interface to clear 510(k). The clearance covers cortical surface recording, monitoring, and stimulation for up to 30-day implant durations.
Clinical status: Cumulative ~37 patients implanted across short-duration studies (intraoperative or up to 30 days), with collaborators including Mount Sinai, Penn, UTHealth, West Virginia, and BIDMC. Note: this is cumulative short-duration use, not permanent implants — important to distinguish from the Neuralink number.
Differentiator: Surface electrodes mean no tissue damage — lowest surgical risk among high-channel-count implants. The 510(k) path is dramatically faster than PMA. The trade-off: surface ECoG signal SNR is lower than penetrating arrays, and whether it can support high-DOF control is the central technical debate.
Blackrock Neurotech & BrainGate — the academic-legacy assets
Blackrock owns the Utah Array — the 96-channel penetrating array used in academic BCI trials for 20+ years. In April 2024, Tether (Tether Evo) took a majority stake for $200M — not a full acquisition, but effectively controlling. MoveAgain System received Breakthrough Designation in 2021, but the pivotal trial timeline keeps slipping.
More importantly, the BrainGate consortium — a collaboration across Brown, MGH, Stanford, Case Western, UC Davis, and others. Cathy Hutchinson (robot arm), Pat Bennett, Ann Johnson — every BCI patient who showed the world what this technology could really do came from BrainGate. A July 2025 medRxiv paper analyzed BrainGate2's 14 participants, 20 implanted arrays, and 2,319 recording sessions — the longest longitudinal dataset in the entire field, and the one FDA reviewers will lean on when evaluating commercial BCI sponsors.
In raw implant-time, Blackrock is still the leader: 30,000 cumulative patient-days, with some recipients running implants for 8+ years. The problem is converting that to commercial product.
Onward Medical — the only company with real revenue from neural interfaces
Listed in Amsterdam. Strictly speaking, not a brain-implant company — but it runs the ARC-BCI program with France's CEA Clinatec, combining a WIMAGINE cortical implant with epidural spinal stimulation as a "digital bridge."
Product line:
- ARC-EX (non-invasive transcutaneous spinal stimulator): FDA de novo authorization in December 2024 (the first non-invasive spinal cord stimulator approved for SCI), CE Mark in Q3 2025, 510(k) for home use in November 2025.
- ARC-IM (implantable spinal stimulator): pivotal trials in Europe and Switzerland.
- ARC-BCI (closed-loop brain + spine): research stage.
Commercial status: ARC-EX is in 60+ US clinics as of Q3 2025, with 40+ systems sold; first EU commercial sales in Q4 2025. This is the only company on the list with meaningful revenue from a neural-interface product.
Europe
INBRAIN Neuroelectronics (Spain / UK)
Founded 2019, spun out of ICN2 Barcelona and the EU Graphene Flagship. Cumulative ~50M Series B in 2024. Core technology is graphene thin-film electrode arrays — high charge injection capacity, excellent biocompatibility, and the ability to do recording and stimulation on the same channel.
September 2024: First-in-human graphene BCI procedure at Salford Royal Hospital in Manchester (during brain tumor resection). First cohort of 4 patients showed no device-related adverse events; the array distinguished healthy from cancerous tissue at micrometer resolution.
CorTec (Germany)
Founded 2010 in Freiburg. Brain Interchange — fully wireless, closed-loop, ECoG electrodes, simultaneous recording and stimulation.
- July 2025: First German-developed BCI implanted in a human — at Harborview Medical Center in Seattle, under an FDA IDE led by the University of Washington. Second patient implanted in February 2026. Indication: stroke motor rehabilitation.
- April 2026: FDA Breakthrough Device Designation — the first BCI ever designated specifically for stroke motor rehabilitation.
Stroke recovery is a much larger commercial indication than ALS: ~800,000 new strokes per year in the US, vs. ~30,000 cumulative ALS patients.
China camp: state-coordinated, semi-invasive bias
NeuroXess — the volume leader
Founded 2021 in Shanghai with deep ties to the CAS Shanghai Institute for Brain Science. Funding is comparatively small: Crunchbase shows ~$29M cumulative (Shanda Group, Sequoia China leading), with likely additional undisclosed domestic rounds.
Core technology is flexible ultra-thin electrodes (~1% of a human hair in diameter), minimally invasive, 64 channels, fully implantable with integrated battery.
Late 2025 milestone: Completed China's first fully implanted, battery-integrated BCI surgery in a 28-year-old man with high cervical SCI (paralyzed for 8 years). Within 5 days he was controlling a cursor, then a wheelchair and robotic arm via brain control.
The biggest open question: NeuroXess claims 54 cumulative human BCI implants since 2022 — which would put them ahead of any single US company and rewrite the leaderboard. This number currently has only single-source Chinese press reporting, with no peer-reviewed paper or independent registry confirmation. Yicai Global and Euronews both reported it, both citing the company. One of the most consequential BCI news items of the next year is whether this number turns out to be real.
NeuCyber NeuroTech / Beinao-1 — the state team
Beinao-1 is a semi-invasive wireless brain chip incubated at the Chinese Institute for Brain Research (CIBR), placed epidurally — between penetrating arrays and non-invasive EEG.
February–March 2025: First 3 human implants completed. Patients with cerebral palsy and ALS used the device to control robotic arms, pour water, and decode ~100 common Chinese words. Target: 13 implants by end of 2025; ~50-patient formal clinical trial in 2026 (pending NMPA approval). Beinao-2 (next-gen with higher signal quality) enters clinical trials in 2026.
The NeuCyber team has publicly admitted they trail Neuralink by ~3 years — unusually candid for Chinese tech narratives.
StairMed
Founded 2021 in Shanghai with Fudan University ties. Cumulative ~$73M raised — Hongshan (Sequoia China), Lilly Asia Ventures, BlueRun, Legend Capital.
The device is a 26mm-diameter, <6mm-thick disk with ultra-flexible electrodes (~1/100 of a human hair). Demonstrated an amputee patient using the BCI for cursor control, chess, and racing games. Commercial launch target: 2028, with the company publicly admitting a 3-year lag versus Neuralink.
Dual-track strategy: DBS (an established revenue category) + motor BCI.
BrainCo — the only one actually shipping
Founded 2015 by Han Bicheng out of Harvard Innovation Lab. HQ Hangzhou + Somerville, MA. Closed ~$286M equity round in January 2026 — the largest single BCI round in China and the second-largest globally after Neuralink's Series E.
Technology is non-invasive EEG + prosthetics. The smart bionic hand sells for RMB 100,000 ($13.8K), 1/5 to 1/7 the price of imported alternatives. Claims 6,000+ amputees served in 2023 alone, with a 30K/year target. Also sells school attention-training headsets, meditation devices, and autism therapy products.
BrainCo is the only company on this list with meaningful product revenue (excluding Onward's spinal stimulator). But non-invasive has a bandwidth ceiling — it cannot compete with invasive BCIs on high-DOF applications.
China state policy (2025)
China is organizing the entire industry top-down — the opposite of the US individual-IDE-driven approach:
- March 2025: The National Healthcare Security Administration created a dedicated insurance category for BCI. Hubei Province set a reference price of RMB 6,552 (~$900) per invasive BCI implant procedure — a bomb under the US business model. If China commercializes BCI implants as cheap reimbursed medical services while US companies are still pricing them as premium devices, the entire cost structure gets redefined.
- July 2025: NMPA released China's first national standard for non-invasive BCI medical devices.
- August 2025: Seven Chinese ministries (MIIT, NDRC, NHC, SASAC, CAS, NMPA, MoST) jointly issued "Implementation Opinions on Promoting the Innovative Development of the BCI Industry" — targeting key breakthroughs by 2027 and industrial leadership by 2030.
Patient-count leaderboard (Q1 2026, cumulative)
| Rank | Company | Approach | Cumulative humans | Notes |
|---|---|---|---|---|
| 1 | NeuroXess (CN) | Flexible intracortical | 54 claimed | Unverified |
| 2 | Precision (US) | Surface thin-film | 37 | Mostly short-duration / intraop |
| 3 | Neuralink (US) | Penetrating intracortical | ~21 | Permanent, FDA IDE |
| 4 | Synchron (US/AU) | Endovascular stent | ~10 | Permanent |
| 5 | Beinao-1 (CN) | Epidural semi-invasive | 3 confirmed | Targeting 13 by year-end |
| 6 | Paradromics (US) | Intracortical microwires | 1 intraop | IDE starts Q1 2026 |
| 7 | CorTec (DE) | Wireless ECoG | 2 | Harborview |
Filter for permanent intracortical implants with multi-month longitudinal data, and Neuralink leads, with Synchron close behind.
Trade-offs across the four electrode philosophies in 2026
| Dimension | Penetrating intracortical | Surface / epidural | Endovascular | Non-invasive |
|---|---|---|---|---|
| Signal quality | Highest (single-unit) | High (LFP/ECoG) | Medium-low | Lowest (EEG) |
| Surgical risk | Craniotomy + thread migration | Mini-craniotomy / burr hole | Interventional radiology, no open surgery | Zero |
| Bandwidth ceiling | 200+ bits/s | Tens | <10 | Tens |
| Regulatory path | Hardest (PMA) | Faster (510(k) possible) | Faster (cardiovascular reviewers) | Quickest |
| Commercial scalability | Bottlenecked on neurosurgeon supply | Moderate | High (interventional radiologists abundant) | Highest |
| Right indication | High-DOF motor, speech, vision | Stroke rehab, tumor mapping | ALS, severe paralysis | Consumer, rehab, prosthetics |
The most interesting shift in 2025–2026: the "intracortical supremacy" narrative is eroding. Precision's 510(k) clearance for surface electrodes and Synchron's Apple/NVIDIA partnerships both signal that regulators and ecosystem partners are willing to accept lower bandwidth in exchange for reduced surgical risk and faster scale-up. Paradromics and Neuralink remain committed to "high-bandwidth intracortical is the only real thing" — that's the central technical contest of the next three years.
Is China actually catching up?
Yes, with asymmetries:
- Execution speed: Beinao-1 went from monkey to 3 human implants in roughly 18 months. That's essentially impossible under an FDA IDE pathway. Even if the NeuroXess 54-implant claim is half true, it still exceeds any single US company.
- Funding scale: The US side still dominates in dollars. Neuralink (1B) + Paradromics + Precision is north of 73M, NeuCyber state-funded, NeuroXess 280M is a signal: capital markets are now willing to write big checks.
- Clinical rigor: US trials go through IDE with serious data integrity standards. Chinese trials are overseen by NMPA, but reporting transparency is lower; patient counts are hard to verify independently. Expect the first wave of peer-reviewed Chinese BCI papers in 2026–2027.
- Policy tailwind: Seven-ministry industrial plan, national standards, dedicated insurance category, the ~$900 reference price in Hubei — these are structural advantages. The US has no equivalent coordinated industrial policy for BCI. The FDA Breakthrough program is the closest analog.
Bottom line: China is unlikely to "beat" Neuralink to a US commercial launch, but it is plausibly the first country where BCI implants will be reimbursed by national insurance and priced for volume, possibly by 2027. That's a different kind of lead.
Realistic 2026–2028 timeline
- 2026: Neuralink releases 3x-capability N1; Synchron starts pivotal trial; Paradromics Connect-One first implants; Beinao-1 starts ~50-patient formal trial in China; Precision tries to expand 510(k) to longer durations.
- 2027: First FDA pivotal trial readouts possible; first NMPA Class III BCI approvals possible.
- 2028: Earliest possible US BCI commercial launch (Synchron is the leading candidate based on publicly committed timelines); Blackrock Neuralace (10,000-channel next-gen) targeting first-in-human demonstration.
Six forward-looking calls for readers
-
2026 will be the year of pivotal trial announcements. Synchron is committed; Neuralink's PRIME/CONVOY may be re-scoped into a registrational pathway; Paradromics Connect-One is a bridge study; Precision will likely try to expand its 510(k) indication to longer durations.
-
Big Tech partnerships will matter more than hardware improvements. The single most strategic move Synchron made in 2025 was native Apple Vision Pro / iPhone integration — that turns a paralysis device into a platform. Expect Neuralink to counter with an Android or xAI integration story; expect Precision and Paradromics to seek OS-level partners.
-
Speech decoding is the real killer app of 2026. UC Davis BrainGate's 2024 speech decoder paper + Paradromics' speech-restoration IDE + Beinao-1's Chinese speech decoding all signal one thing: BCI has moved past "cursor control" as the benchmark. The next wow demo will be a locked-in ALS patient holding a real-time natural conversation.
-
Watch device-longevity disclosures. The Achilles' heel of intracortical BCIs is thread retraction and signal degradation. Neuralink publicly acknowledged this in patient one. The BrainGate 14-participant longevity paper (July 2025) is the best published dataset. Any company that can demonstrate multi-year stable high-performance recordings will own a massive competitive moat.
-
China's $900 reimbursement price is a bomb. If China sells BCI implants as cheap reimbursed medical services while the US still treats them as premium devices, the entire cost stack comes under pressure. Expect US companies to start talking reimbursement during pivotal trial design, not after launch.
-
Geopolitical risk is rising. Both the US and China treat BCI as a strategic technology. Expect export controls on BCI-adjacent chips and cross-border investment scrutiny to tighten in 2026–2027. Tether's stake in Blackrock Neurotech and NEOM's investment in Paradromics are early canaries.
Open questions I'm tracking
- Independent verification of NeuroXess's 54 implants. If China's ClinicalTrials.gov-equivalent has the registry entries, the leaderboard rewrites.
- Neuralink thread retraction status. How many of the 21 participants have thread issues, and what's the mitigation? No public disclosure.
- Synchron pivotal trial design. What's the endpoint? What's N? PMA or de novo?
- First NMPA BCI implant approval in China — StairMed (2028 target), NeuCyber (state team), or NeuroXess (fastest execution)?
- Paradromics Connect-One readout timing.
- Onward ARC-BCI next data drop. The brain-spine digital bridge is the most scientifically elegant demo in the field — when does the next clinical readout land?
BCI is no longer a "next decade" story — it's now a "next three years, who lands FDA approval and who gets covered by Chinese insurance" story. How would you place the bet?
References
Neuralink
- PRIME Study Progress Update
- Neuralink raises $650M Series E
- Neuralink expands BCI trial to 21 participants (Jan 2026)
- ClinicalTrials.gov: PRIME (NCT06429735)
- IEEE Spectrum on Blindsight
Synchron
- COMMAND Study positive safety results (Sep 2024)
- [Synchron 200-Million-Series-D-to-Advance-Brain-Computer-Interface-Technology)
- Synchron × Apple native BCI-HID integration (May 2025)
- Synchron × NVIDIA Holoscan + Vision Pro demo
Paradromics
- Connect-One FDA IDE approval (Stat News, Nov 2025)
- First-in-human recording (May 2025)
- Citeline coverage
Precision Neuroscience
- FDA 510(k) clearance press release (Apr 2025)
- FDA 510(k) document K242618 (PDF)
- MedTech Dive coverage
Blackrock Neurotech & BrainGate
- Tether takes majority stake in Blackrock Neurotech
- Blackrock MoveAgain Breakthrough Designation
- BrainGate 14-participant longevity preprint (Jul 2025)
- BrainGate clinical trials portal
Onward Medical
INBRAIN Neuroelectronics
CorTec
China camp
- NeuroXess first battery-integrated BCI implant (Yicai Global)
- Euronews: paralysed man controls devices in five days
- Beinao-1 first human implants (Yicai Global)
- StairMed $73M raise
- BrainCo ~$286M equity round (Yicai Global)
- BrainCo $280M raise analysis (NeuroFounders)
- SCMP: China targets $900/procedure BCI pricing
- China BCI seven-ministry plan (CSET Georgetown analysis)
- Global Times BCI coverage
- CNN: China brain tech accelerates
- TheDebrief: 50 human implants + Chinese speech decoding
