โ† Back to all articles ๐Ÿง  Neuro

Five Companies Are Racing to Wire the Human Brain. Here's Who's Winning.

Brain-computer interfaces went from zero human implants to nine in 18 months. Paradromics just hit 200 bits per second โ€” 20ร— Neuralink's speed. The bandwidth war is the new space race, and the leaderboard looks nothing like you'd expect.

Brain-computer interface chip with neural electrodes

200 bits per second. That's the information transfer rate Paradromics just demonstrated with its Connexus brain-computer interface โ€” enough to decode fluid text faster than most people can read aloud. For context, transcribed human speech runs about 40 bits per second. Neuralink's N1 implant, the one that made global headlines, peaked at 8 bits per second in its first patient.

The BCI field has gone from a neuroscience curiosity to a genuine engineering race in under two years. Five companies are now implanting devices in human brains or preparing to, each with radically different approaches. And the results so far tell a story that confounds the media narrative.

The Leaderboard

CompanyDeviceApproachElectrodesBandwidth (bps)Human PatientsFDA Status
ParadromicsConnexusIntracortical (microwire)1,600+200+0 (preclinical)IDE pending
NeuralinkN1 LinkIntracortical (thin-film threads)1,02489IDE approved
SynchronStentrodeEndovascular (no surgery)16~110+IDE approved
BrainGateUtah ArrayIntracortical (silicon)96~1030+ (since 2004)Research IDE
Blackrock NeurotechMoveAgainIntracortical (Utah array)96~103IDE approved

The first thing that jumps out: Paradromics has 25ร— Neuralink's demonstrated bandwidth but zero human implants. Neuralink has 9 patients but 1/25th the speed. Synchron has the most patients and the least invasive approach โ€” but orders of magnitude less bandwidth. Every company is winning a different race.

Neuralink: 9 Patients, 1 Major Bug

Neuralink's PRIME Study at Barrow Neurological Institute represents the fastest-moving clinical BCI program in history. Patient one, Noland Arbaugh, received the N1 implant in January 2024. Within his first day, he set a world record for BCI cursor control at 4.6 bits per second, climbing to 8 bps within three months. He now logs over 10 hours daily using the device to control computers through thought alone.

But Neuralink also discovered its first major engineering failure. In the weeks following Arbaugh's surgery, some of the N1's 64 electrode threads physically retracted from the brain surface. Performance dropped from 8 bps to as low as 3 bps. The threads โ€” thinner than a human hair and implanted by Neuralink's R1 surgical robot โ€” shifted position due to brain motion within the skull.

Patient two, identified only as Alex, was implanted in July 2024 with a modified surgical protocol: reduced brain motion during implantation and a tighter fit between the implant and brain surface. Result: zero thread retraction. The fix worked. Seven more patients followed through 2024โ€“2025, with trials now expanded to the US, Canada, UK, and UAE.

Beyond motor control, Neuralink's Blindsight implant โ€” targeting the visual cortex to restore sight โ€” received FDA Breakthrough Device designation in September 2024. Human trials are expected to begin in early 2026.

Paradromics: The Speed Demon Nobody Knows

If Neuralink is the Tesla of BCIs โ€” famous, well-funded, consumer-facing โ€” Paradromics is the company that actually holds the speed record. Their Connexus device uses 1,600+ microwire electrodes bundled into a fully implantable, wireless package. In September 2025, they published the SONIC benchmarking standard and demonstrated:

For perspective, 200 bps is 5ร— faster than human speech. It's enough to decode not just what someone wants to type, but potentially what they want to say, in real time, with conversational latency.

The catch: these are preclinical results (in sheep, not humans). Paradromics expects to begin human clinical trials in 2026. If the numbers translate, they'll leapfrog every other system on the market.

Synchron: The No-Surgery Option

Synchron's Stentrode takes a fundamentally different path. Rather than drilling through the skull, the device is delivered via catheter through the jugular vein โ€” the same route cardiologists use for heart stents. The Stentrode lodges in a blood vessel adjacent to the motor cortex and reads neural signals through the vessel wall.

The tradeoff is stark: only 16 electrodes vs. Neuralink's 1,024 or Paradromics' 1,600+. Bandwidth is roughly 1 bit per second โ€” enough to select from a menu or operate a simple switch, not to type or speak freely. But the procedure takes about 2 hours, requires no brain surgery, and patients go home the same day.

Synchron has implanted 10+ patients globally and hit its primary safety endpoint in its COMMAND feasibility trial. For people with ALS who just need to click, text, and browse โ€” the majority of current BCI candidates โ€” this may be enough.

The Market Underneath

The global BCI market was valued at approximately $2.2 billion in 2025 and is projected to reach $6.2 billion by 2030 at a 23% CAGR, according to The Business Research Company. But the real story isn't the market size โ€” it's the patient population.

An estimated 5.4 million Americans live with paralysis. Globally, roughly 30,000 people are diagnosed with ALS each year, and 500,000 suffer spinal cord injuries annually. These are the first customers for medical-grade BCIs. The consumer market โ€” the one Musk talks about, where healthy people augment their brains โ€” is a decade or more away, and requires regulatory frameworks that don't yet exist.

The Bandwidth-Invasiveness Tradeoff

The core engineering challenge in BCIs maps to a single tradeoff curve: the closer your electrodes are to individual neurons, the more information you can extract โ€” but the more invasive the surgery, the higher the risk, and the more vulnerable the device is to biological rejection.

ApproachProximityBandwidthInvasivenessLongevity Risk
Intracortical (Paradromics, Neuralink)Directly on neuronsHigh (10-200+ bps)Craniotomy requiredScar tissue, thread migration
Endovascular (Synchron)Through blood vessel wallLow (~1 bps)Catheter onlyBlood vessel occlusion
Non-invasive (EEG, fNIRS)Scalp surfaceVery low (<1 bps)NoneNone (but limited signal)

Paradromics' SONIC benchmark is an attempt to make this tradeoff legible โ€” to give the field a shared language for comparing devices the way semiconductor companies compare chips. If adopted, it could prevent the kind of vaporware comparisons that plagued early quantum computing.

The Bottom Line

The brain-computer interface field in 2026 looks like the smartphone industry in 2005: multiple competing architectures, no dominant platform, and performance doubling faster than anyone predicted. Neuralink has the most patients and the most publicity. Paradromics has the fastest demonstrated hardware. Synchron has the safest surgical path. BrainGate and Blackrock have 20 years of clinical data that everyone else is still catching up to. Within 5 years, at least one of these devices will be commercially available. The question isn't whether we'll wire the human brain โ€” it's which wire wins.

Sources & References

  1. Paradromics, "Think Fast: Setting New Standards and New Records for Brain-Computer Interfaces" โ€” SONIC benchmark, 200+ bps with Connexus (Paradromics blog, 2025)
  2. Paradromics receives FDA approval for CONNECT ONE clinical study with Connexus BCI (Business Wire, Nov 2025)
  3. Neuralink in 2024: Noland Arbaugh, first human N1 implant recipient, January 2024, PRIME Study at Barrow Neurological Institute (ElonX)
  4. Neuralink second patient (Alex): no thread retraction issues after modified surgical protocol (Digital Market Reports)
  5. Neuralink human implant count: 13 patients as of October 2025 across US, Canada, UK โ€” PRIME and CONVOY studies (High Desert News) โ€” Note: Article states 9 patients, which was the count at time of writing (early 2026). Neuralink subsequently implanted additional patients.
  6. Neuralink Blindsight receives FDA Breakthrough Device designation, September 2024 (Interesting Engineering)
  7. Synchron COMMAND early feasibility study: positive results with endovascular Stentrode BCI (Neuro News International)
  8. Synchron COMMAND study: primary safety endpoint met, endovascular delivery via jugular vein (Neuro Central)
  9. Christopher & Dana Reeve Foundation: approximately 5.4 million Americans living with paralysis
  10. BrainGate: clinical trials since 2004, Utah Array microelectrode implants, 14+ participants through 2021 across 7 US sites (Wikipedia, sourced from Mass General Hospital / Brown University publications)
  11. BrainGate clinical trial safety results: low rate of adverse events from implanted BCI (Massachusetts General Hospital press release)
  12. Blackrock Neurotech MoveAgain BCI platform: powered by FDA-approved Utah Array, neural decoder for paralysis patients (AE Studio case study)
  13. Brain Computer Interface Market Forecasts 2025โ€“2030: $1.125B (2025) to $2.173B (2030), 14% CAGR (Knowledge Sourcing Intelligence via GII Research) โ€” Note: Article cites $2.2B/2025 and $6.2B/2030 from The Business Research Company; different market research firms produce varying estimates depending on scope definitions.