# Is the BCI Industry Finally Hitting Commercial Scale in 2026?

China's [brain-computer interface](https://bciintel.com/glossary/brain-computer-interface) sector recorded 64 investment and financing transactions totaling approximately 7.253 billion yuan through June 30, 2026 — already surpassing the full-year 2025 total, according to data cited by 36Kr from IT Juzi. That capital surge is arriving alongside two significant technical milestones: [Neuralink Corp](https://bciintel.com/companies/neuralink)'s July 1 announcement of a transdural electrode implantation technique that bypasses conventional dura incision, and Meta's June 2026 release of the Brain2Qwerty v2 model in collaboration with Spain's Basque Cognitive Center, which reportedly raised non-invasive decoding accuracy from 8% to 61%. Together, these data points suggest the industry is transitioning from proof-of-concept neuroscience to capital-intensive industrialization — though the distinction between early feasibility milestones and commercially validated products remains critical for investors and clinicians alike.

The single largest disclosed transaction in China was [BrainCo](https://bciintel.com/companies/brainco)'s 2 billion yuan financing round announced earlier this year. BrainCo has also previewed what it describes as the world's first brain-controlled robot training platform, a development that intersects BCI with humanoid and assistive robotics — a convergence being tracked closely at [humanoidintel.ai](https://humanoidintel.ai).

---

## Neuralink's Transdural Approach: What Changed on July 1

On July 1, [Neuralink Corp](https://bciintel.com/companies/neuralink) announced completion of a surgical procedure it characterizes as the world's first intracortical BCI implantation that penetrates the dura mater without cutting it — threading electrode wires directly through the dural membrane rather than opening it conventionally. The clinical and engineering rationale is straightforward: dural incision creates a larger breach point for cerebrospinal fluid leakage, infection risk, and inflammatory response. A transdural puncture approach, if it holds up across more subjects, could meaningfully reduce surgical trauma and shorten recovery timelines.

**What the source does not tell us — and what matters:**

The 36Kr report does not specify how many subjects have undergone this procedure, whether this was performed under Neuralink's existing IDE (Investigational Device Exemption) with the FDA, or what electrode count and array geometry were used. It also does not report functional outcomes — decoding accuracy, bits per second throughput, or adverse event data. Until Neuralink publishes peer-reviewed data or files updated IDE documentation with specific procedural outcomes, this announcement should be treated as a company-disclosed surgical technique update, not a validated clinical result.

Nonetheless, the technical direction is meaningful. Long-term biocompatibility and foreign body response at the dura-electrode interface has been a persistent engineering challenge across the intracortical field. If transdural insertion reduces the mechanical disruption footprint, it addresses a real obstacle to scaling implant procedures beyond specialized neurosurgical centers.

---

## Meta Brain2Qwerty v2: Non-Invasive Decoding Hits 61%

The more technically provocative data point in the 36Kr report is Meta's Brain2Qwerty v2 result: a jump in non-invasive [EEG](https://bciintel.com/glossary/eeg) decoding accuracy from 8% to 61%, achieved in collaboration with Spain's Basque Cognitive Center and released in June 2026. If independently validated, this represents a substantial compression of the performance gap between non-invasive and invasive approaches for communication applications.

The physics of the problem is unforgiving: the skull attenuates more than 80% of neural electrical signals, leaving surface EEG operating in the microvolt range with high noise floors. The 36Kr analysis frames this accurately — decoding motor intention, emotional state, or typed characters from scalp EEG is an engineering challenge comparable in difficulty to the surgical precision required for intracortical implantation. The improvement in Brain2Qwerty v2 presumably reflects architectural advances in the neural decoding model itself, though the source does not detail the model's parameter count, training data volume, or the specific experimental paradigm (imagined speech, attempted movement, visual evoked potentials, etc.).

**The skeptical read:** A jump from 8% to 61% accuracy in a single model iteration is large enough to warrant scrutiny. Key questions for any follow-up analysis: accuracy on what task, under what conditions, with how many subjects, and whether the 61% figure represents online decoding or offline post-hoc analysis of recorded data. Offline accuracy figures routinely outperform online BCI operation by substantial margins — a distinction that has tripped up multiple press cycles in this field.

---

## China's Three-Track BCI Investment Structure

The 36Kr analysis describes China's BCI investment landscape as organized around three parallel development routes, each with a distinct commercial logic:

**Invasive route** — High signal fidelity, direct cortical access, highest regulatory burden, longest clinical verification cycle. Suited for long-horizon institutional and strategic capital targeting severe neurological diseases. The signal quality advantage is real: intracortical arrays record individual action potentials and local field potentials unavailable to surface methods, which is why this route remains the primary technical path for restoring communication and motor function in conditions like ALS and tetraplegia.

**Non-invasive route** — Already generating commercial revenue in education (attention monitoring), consumer electronics, and human-computer interaction. [BrainCo](https://bciintel.com/companies/brainco) is the most visible Chinese player here, citing independently developed solid-state gel electrodes that achieve a signal-to-noise ratio of 10.78 dB in real EEG tests and motor intention recognition latency of under 200 milliseconds. [Dry electrodes](https://bciintel.com/glossary/dry-electrode) eliminating conductive gel are a genuine usability advance for consumer and educational deployment, even if the underlying signal quality remains inferior to wet or implanted approaches.

**Semi-invasive route** (ECoG / [electrocorticography](https://bciintel.com/glossary/electrocorticography)) — Epidural or subdural electrode arrays that offer better signal quality than scalp EEG without full intracortical penetration. Regulatory and surgical complexity falls between the other two routes. Companies like [Synchron](https://bciintel.com/companies/synchron) (endovascular approach) and [Precision Neuroscience](https://bciintel.com/companies/precision-neuroscience) (surface micro-electrode array) occupy analogous positions in Western markets.

[Merge Labs](https://bciintel.com/companies/merge-labs), backed by Sam Altman, is cited in the report as having bet on the non-invasive route, positioning it as a counterpart to Neuralink's invasive approach in the global investment narrative.

---

## BrainCo's Closed-Loop Haptics and the Bidirectional BCI Race

Beyond decoding, BrainCo claims a two-way [closed-loop](https://bciintel.com/glossary/closed-loop) system for haptic feedback: flexible tactile sensors identify object features, convert contact information into micro-stimuli perceivable by the user, and deliver sensory feedback through neural and electromyographic pathways. This bidirectional architecture — outbound motor intent decoding plus inbound somatosensory feedback — is what separates a prosthetic controller from a genuine neuroprosthetic system. The source does not provide peer-reviewed validation data for this claim, so it should be read as a company-reported capability pending independent verification.

---

## Industry Trajectory and Clinical Translation Timeline

The 7.253 billion yuan H1 2026 figure for China alone, measured against a global backdrop that includes Neuralink's ongoing U.S. clinical trial, Synchron's IDE-approved COMMAND trial, and Precision Neuroscience's Layer 7 Cortical Interface feasibility work, suggests the industry is entering a phase where capital availability is no longer the primary bottleneck. The bottlenecks that remain are: regulatory pathway clarity for semi-invasive and invasive devices in China and the EU; long-term implant stability data beyond two years; and the fundamental challenge of demonstrating clinical benefit in controlled trials, not just feasibility studies.

For patient access, the near-term picture is clearest in non-invasive applications — assistive communication tools, attention-regulation systems for education, and prosthetic control interfaces are already in commercial deployment in China and select Western markets. Invasive systems capable of restoring meaningful motor or communication function to patients with ALS or tetraplegia remain in IDE-stage clinical trials globally, with commercial approval trajectories measured in years, not months.

---

## Key Takeaways

- China's BCI sector recorded **64 deals worth approximately 7.253 billion yuan** through June 30, 2026, surpassing all of 2025 per IT Juzi data cited by 36Kr.
- **BrainCo raised 2 billion yuan** in the largest disclosed Chinese BCI transaction of the period.
- **Neuralink announced a transdural implantation technique** on July 1 that threads electrode wires through the dura without incision, aiming to reduce surgical trauma — but no functional outcome data has been published.
- **Meta's Brain2Qwerty v2** reportedly raised non-invasive EEG decoding accuracy from 8% to 61% in collaboration with Spain's Basque Cognitive Center — a figure that requires independent replication and task-condition disclosure.
- China's BCI investment follows three tracks: invasive (long-cycle, high-regulatory), non-invasive (commercial-ready), and semi-invasive (intermediate risk/return).
- The performance gap between non-invasive and invasive approaches is narrowing on paper, but intracortical systems retain fundamental signal quality advantages for severe neurological applications.
- Capital abundance is no longer the primary constraint — regulatory clarity, long-term stability data, and controlled trial evidence are.

---

## Frequently Asked Questions

**How much has China invested in BCI in 2026?**
According to IT Juzi data cited by 36Kr, China's BCI sector recorded 64 investment and financing transactions totaling approximately 7.253 billion yuan through June 30, 2026, already exceeding the full-year 2025 total.

**What did Neuralink announce on July 1, 2026?**
Neuralink announced completion of a brain-computer interface implantation surgery that threads electrode wires through the dura mater without cutting it — a transdural approach intended to reduce surgical trauma and recovery time compared to conventional dural incision. No peer-reviewed outcome data has been published.

**What is Meta's Brain2Qwerty v2 and how accurate is it?**
Brain2Qwerty v2 is a non-invasive BCI decoding model developed by Meta in collaboration with Spain's Basque Cognitive Center, released in June 2026. The 36Kr report states it raised decoding accuracy from 8% to 61%. Task conditions, subject count, and whether this reflects online or offline decoding have not been specified in the source material.

**What is the difference between invasive and non-invasive BCIs in terms of investment logic?**
Invasive BCIs require surgical implantation, face longer regulatory pathways, and target severe neurological conditions — making them suited for long-horizon capital. Non-invasive BCIs using scalp EEG are already commercially deployed in education and consumer electronics, offering lower risk and clearer near-term revenue but lower signal fidelity.

**When will invasive BCIs be commercially available?**
Invasive BCI systems are currently in IDE-stage clinical trials globally (including Neuralink's PRIME study in the U.S.). Commercial approval for intracortical systems targeting ALS or tetraplegia is measured in years, dependent on controlled trial results, long-term safety data, and regulatory pathway decisions in relevant jurisdictions. No commercially approved invasive BCI for motor or communication restoration currently exists outside of specific compassionate use contexts.

---

*This article is based on a single Chinese-language industry analysis published by 36Kr on July 8, 2026. Specific figures (7.253 billion yuan, 64 transactions, 10.78 dB SNR, 200ms latency, 61% decoding accuracy) are sourced from that report and have not been independently verified by bciintel.com. Results cited for Neuralink and BrainCo derive from company announcements, not peer-reviewed publications. Nothing in this article constitutes medical advice. All clinical findings referenced are from early-stage feasibility work unless otherwise noted.*