Draft, not yet released. In stewardship review. Not indexed until the first approved release.

Executive summary

The edition on a few pages: the abstract, the key messages, the headline finding and what each steward reads in it.

7 min read1 figure0 tables15 notesNR-XTECH-NEURO-GLOBAL-2026-BASELINE

Contents
  1. ·Edition home
  2. ·Executive summary
    1. S.1 Abstract
    2. S.2 Key messages
    3. S.3 Headline finding
    4. S.4 At a glance
    5. S.5 What each steward reads in it
    6. S.6 What this edition is, and is not
  3. 1Neurotechnology is clinical hardware and a new class of personal data
  4. 2Key findings
  5. 3Systems and sectors at baseline
  6. 4Risks and capability
  7. 5Consumer neural data outside medical rules
  8. 6Readings from the other desks
  9. 7The 2027 agenda
  10. 8Systems in depth
  11. AMethodology
  12. BData appendix
  13. CReferences and record
  14. ·Complete edition, for print

S.1Abstract

Neurotechnology matters to the Nexus programme in two ways. It is a new class of implanted hardware, built from electrode arrays, amplifiers, radios and software decoders that need materials, power, surgeons and long-term follow-up. It is also a new source of personal data, because the signals of the nervous system can reveal intent, and sometimes inner speech. The clinical evidence is strong for single participants. A speech decoder reached 62 words per minute on a 125,000-word vocabulary1, and another reached a median of 78 words per minute2. A later study reported 97.5% accuracy beyond eight months after implantation, in one participant3. A fully implanted brain-spine interface stayed stable for over one year7, and an endovascular array recorded for one year in five participants8. Money and rules are moving. Neuralink reported five implanted participants and a USD 650 million round in 202520. China published an implementation opinion with targets for 2027 and 203016. UNESCO adopted a Recommendation on the Ethics of Neurotechnology in November 202510, Colorado extended its privacy law to neural data in 202414, and a US Senate bill would direct the FTC to study neural data11. The physical footprint, the workforce and market size are not measured on a common basis. This baseline sets out what the evidence supports and where it is thin. It forecasts no capability date and ranks no country or company.

S.2Key messages

Six readings that frame the neurotechnology baseline, and the eight questions this edition answers, each with its status and source.

  1. In single participants, at 62 to 78 words per minute and up to 97.5% accuracy.123

    Do speech interfaces work?Reported

  2. One interface was stable over one year and an endovascular array recorded for one year.78

    Can an implant last?Reported

  3. Imagined sentences were decoded in four participants; safeguards are proposed.5

    Can inner speech be read?Reported

  4. UNESCO adopted a Recommendation; Colorado covers neural data; a US Senate bill is in committee.101411

    What do the rules say?Reported

  5. A 2021 constitutional rule and a 2023 deletion order; inferred data remain a grey zone.1718

    What does Chile show?Reported

  6. A USD 650 million company round in 2025; market size is not on a common basis.2029

    What does the money look like?Reported

  7. Not bound in this edition.9

    What is the power and heat footprint?Not bound

  8. Consumer neural data outside medical rules while clinical interfaces advance. SCN-XT20-NDT-01

    What will 2027 test first?Expert input

S.3Headline finding

ReportedEQL 4NRF-XTECH-NEURO-GLOBAL-2026-BASELINE-INF-001

Two Nature papers of August 2023 reported speech decoding at 62 and 78 words per minute in single participants with paralysis.

Figure S.1Speech decoding rates and natural conversation (words per minute)
ItemBar, scaled to the largest valueValue
Willett 2023, 125,000-word vocabulary62
Metzger 2023, median rate78
Natural conversation, as stated in the first paper160

Sources: Nature 620 (2023)12. The 160 figure is the approximate rate of natural conversation as stated in1.

Decoding of attempted speech into text is now fast enough to support conversation in a laboratory setting for a few people. The first study reported a 23.8% word error rate on a large vocabulary and the second a median of 25%; a later study in one participant reported 97.5% accuracy beyond eight months3. Each result rests on one participant and a research system.

S.4At a glance

62words per minute

speech decoding, 20231

97.5%

accuracy beyond eight months3

5participants

one-year endovascular study8

650USD million

company round, 202520

194Member States

entrusted UNESCO mandate in 202310

2030target year

Chinese implementation opinion16

S.5What each steward reads in it

GCRI · KEY T

Technical Truth

The evidence supports reading neurotechnology as clinical hardware and as a source of personal data.

GRF · KEY M

Public Meaning

For most people the neurotechnology story arrives as a headline about a brain implant or a question about mental privacy.

GRA · KEY F

Capital Meaning

Readiness implications only. Rounds are large and rest on company statements, market estimates differ widely and reimbursement is not reviewed.

NEXUS

Synthesis

Neurotechnology is well described by single results and thinly measured as a system.

S.6What this edition is, and is not

What this edition is

A baseline: the 2026 state of neurotechnology as clinical hardware and personal data and the couplings through which it carries risk into other systems. It synthesises published evidence and ranks no country, company or asset.

What it is not

Not a forecast of when any device will reach the market, a technology ranking, an endorsement of any company or an assessment of any programme. Medical-device rules, privacy law and funding are decisions of competent authorities.

What comes next

National, regional and sector evidence feeds the 2027 neurotechnology edition through the federation. Each edition reports what changed against this baseline: observations, source revisions, method changes, scenario updates and corrections.

Notes

Numbers are the edition’s reference numbers, so a note reads the same in every chapter. The full list, with every source and its licence, is in Annex C. Annex C, references

  1. 1Willett, F.R. et al., A high-performance speech neuroprosthesis, Nature 620(7976), 1031 to 1036, published online 23 August 2023 (reference record read: 62 words per minute; 23.8% word error rate on 125,000 words; 9.1% on 50 words). Springer Nature, London. Observation year 2023. Accessed 2026-11-03.back
  2. 2Metzger, S.L. et al., A high-performance neuroprosthesis for speech decoding and avatar control, Nature 620(7976), 1037 to 1046, published online 23 August 2023 (reference record read: median 78 words per minute; median word error rate 25%). Springer Nature, London. Observation year 2023. Accessed 2026-11-03.back
  3. 3Card, N.S. et al., An accurate and rapidly calibrating speech neuroprosthesis, medRxiv preprint, 26 December 2023 (preprint record read: 99.6% on 50 words on day 1; 90.2% on 125,000 words on day 2; 97.5% beyond eight months; 256 electrodes; the record notes the New England Journal of Medicine version 391(7), 609 to 618, August 2024). Cold Spring Harbor Laboratory, New York. Observation year 2023. Accessed 2026-11-03.back
  4. 7Lorach, H. et al., Walking naturally after spinal cord injury using a brain-spine interface, Nature 618(7963), 126 to 133, published online 24 May 2023 (reference record read: fully implanted; stable for over one year; one man with chronic tetraplegia). Springer Nature, London. Observation year 2023. Accessed 2026-11-03.back
  5. 8Chetty, N. et al., Signal properties and stability of a chronically implanted endovascular brain computer interface, medRxiv preprint, 25 September 2025 (preprint record read: five participants; 16-channel array; one year; not peer reviewed). Cold Spring Harbor Laboratory, New York. Observation year 2025. Accessed 2026-11-03.back
  6. 20Zeff, M., Elon Musk Neuralink closes a USD 650 million Series E (trade report, 2 June 2025; page read; vendor statement as reported: five people implanted; valuation of about USD 9 billion pre-money from an earlier press report). TechCrunch, San Francisco. Observation year 2025. Accessed 2026-11-03.back
  7. 16Ministry of Industry and Information Technology and six other bodies (eds.), Implementation Opinions on Promoting the Innovation and Development of the Brain-Computer Interface Industry, MIIT (2025) No. 164, 23 July 2025, English translation by the Center for Security and Emerging Technology, 18 December 2025 (translation page read). Georgetown University, Washington. Observation year 2025. Accessed 2026-11-03.back
  8. 10UNESCO (ed.), Recommendation on the Ethics of Neurotechnology (page read: adopted November 2025 at the 43rd session of the General Conference; mandate entrusted by 194 Member States in 2023; the page states no neural data provisions). United Nations Educational, Scientific and Cultural Organization, Paris. Observation year 2025. Accessed 2026-11-03.back
  9. 14Colorado General Assembly (ed.), HB24-1058 Protect Privacy of Biological Data (bill page read: signed 17 April 2024; effective 7 August 2024). Colorado General Assembly, Denver. Observation year 2024. Accessed 2026-11-03.back
  10. 11United States Senate (ed.), Management of Individuals Neural Data Act of 2025 (MIND Act), S. 2925, 119th Congress, introduced text (text read: FTC study and report within one year of enactment; USD 10 million authorised; referred to committee). US Government Publishing Office, Washington. Observation year 2025. Accessed 2026-11-03.back
  11. 5Kunz, E.M. et al., Inner speech in motor cortex and implications for speech neuroprostheses, Cell 188(17), 4658 to 4673, published online 14 August 2025 (reference record read: four participants). Cell Press, Cambridge MA. Observation year 2025. Accessed 2026-11-03.back
  12. 17Cornejo-Plaza, M.I., Cippitani, R. and Pasquino, V., Chilean Supreme Court ruling on the protection of brain activity: neurorights, personal data protection, and neurodata, Frontiers in Psychology 15, 27 February 2024 (reference record read). Frontiers Media, Lausanne. Observation year 2024. Accessed 2026-11-03.back
  13. 18Lee, B.H., Even Chile neurorights leave inferred mental data in a gray zone, Stanford Law School Law and Biosciences Blog, 27 April 2026 (blog page read; commentary, not a legal text). Stanford University, Stanford. Observation year 2026. Accessed 2026-11-03.back
  14. 29UNVERIFIED, not opened: market-research reports on the neurotechnology and brain-computer interface markets for 2025 (several firms; figures differ by a large factor depending on definition), taken from search-result summaries; weak evidence and no figure used. Observation year 2025. Accessed 2026-11-03.back
  15. 9US Food and Drug Administration (ed.), Implanted Brain-Computer Interface (BCI) Devices for Patients with Paralysis or Amputation: Non-clinical Testing and Clinical Considerations, final guidance, May 2021 (guidance page read). Center for Devices and Radiological Health, Silver Spring. Observation year 2021. Accessed 2026-11-03.back

Cite this summary

Valadbaygi, S. (Ed.). (2026). Executive summary: Executive summary. In Exponential Technology Report 2026: Neurotechnology and Human–Machine Interfaces (Baseline Edition 2026, Report No. NR-XTECH-NEURO-GLOBAL-2026-BASELINE) [Draft under review]. The Global Centre for Risk and Innovation (GCRI).

Every citation form of the whole report, BibTeX and RIS included, is on the edition home. Cite this report

ReportsContribute

Send evidence or a correction on Neurotechnology and Human–Machine Interfaces.

Reports is the Consortium’s quarterly publication, written by six desks. This page sits with the Global Centre for Risk and Innovation (GCRI) Systems Evidence desk. What you know about this subject can become evidence, a case or a correction in the next edition, credited to you and published under CC BY 4.0 where it is your own work. Sent to the wrong desk, it is forwarded, not refused.

GCRI Systems Evidence deskExponential Technology

Never send personal data, sensitive infrastructure detail or anything you have no right to share. A contribution is reviewed before use; sending it confers no authorship or editorial rights, and sponsorship never buys a finding.

CC BY 4.0 Free to read. Text and figures: CC BY 4.0. Data: CC0 1.0. Authors keep copyright. Published by GCRI Canada.

The Global Centre for Risk and Innovation

Science · evidence · measurement

Not for profit

Our articles, maps and twins are free to read and free to republish. Readers and partners keep it that way.

Partner with us

Who funds us

GCRI is a not-for-profit institution. To ask who funds a piece of work, use the funding questions form.

Independence

Editors and the expert reviewers named on each article decide what we publish. Funders and partners do not review or approve articles before publication.

Corrections policy

Licence

Articles: CC BY 4.0. Figures, maps, twins and data: CC BY 4.0. Other uses: ask through the permissions form.

Privacy choices