Get alerts on this situation
We’ll email you as it develops, and you can follow the whole thread from day one.
Unsubscribe anytime.
[SITUATION] · [ACTIVE]
3 clusters · 10 sources · 24 days · First seen · Last updated
Categories: TECHNOLOGY
China‑Korea neuromorphic & photonic chip breakthroughs
Entities: Chinese Academy of Sciences · Peking University · memristor chip · Yuchao Yang · brain‑speed chip
Overview
In July 2026 the Peking University and Chinese Academy of Sciences team released a 40 nm neuromorphic chip that can emulate cortical dynamics with sub‑10 ms latency, delivering up to 478‑times faster reconstruction of brain surfaces than an Nvidia A100 GPU while consuming a fraction of the power. The chip’s in‑memory compute‑in‑memory architecture, based on a conductance‑drift array, was shown to model both white‑matter and gray‑matter structures, supporting real‑time brain‑computer interfaces, surgical navigation and research on neuro‑degenerative diseases.
A joint effort with South Korean researchers yielded a programmable photonic integrated circuit that exploits the coupled‑resonator induced transparency (CRIT) effect to provide on‑chip delay, buffering and frequency conversion, promising energy‑efficient AI acceleration for data‑center workloads.
On 27 July the Chinese group demonstrated the chip’s 0.28 mm² device operating at 50 MHz and confirmed the performance gap (up to several hundred‑fold faster than GPUs) while maintaining low power. The following day a memristor‑based version was reported in *Science*, showing 3.8–36.3 × faster operation than state‑of‑the‑art ASICs and 11.8–24.7 × lower energy use. Both advances reinforce the trajectory toward ultra‑low‑latency, low‑power neuromorphic platforms that integrate photonic support for next‑generation AI and biomedical applications.
Claims
What the coverage asserts, and how well corroborated each claim is across sources.
- [● 4 SOURCES] Researchers at Peking University and the Shanghai Institute of Microsystem and Information Technology developed a memristor chip that can model the human brain surface in under 10 milliseconds. (Peking University, Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences)
- [● 4 SOURCES] The chip performs computations directly in memory, eliminating data transfer between memory and processor. (Developers of the memristor chip)
- [● 4 SOURCES] The chip is fabricated with a 40‑nm manufacturing process and occupies 0.28 mm². (Chip specifications)
- [● 4 SOURCES] In neural dynamics calculations the chip is 3.82‑to‑36.27 times faster than state‑of‑the‑art ASICs. (Performance benchmarks)
- [● 4 SOURCES] The chip consumes 11.75‑to‑24.73 times less power than those ASICs. (Performance benchmarks)
- [● 4 SOURCES] For 3‑D brain‑cortex reconstruction the chip is up to 478 times faster than an NVIDIA A100 GPU. (Performance benchmarks)
- [● 4 SOURCES] Potential applications include brain‑computer interfaces, digital brain twins, real‑time surgical navigation, and research on Alzheimer’s and Parkinson’s diseases.
Timeline
-
3 days ago
[TECHNOLOGY] 4 sourcesPeking University memristor chip models brain in under 10 msA Chinese memristor chip developed by Peking University can model brain folds in under 10 ms, running 3.8–36.3 × faster than ASICs, using far less power, and may aid brain‑computer interfaces and disease study.
-
5 days ago
[TECHNOLOGY] 3 sourcesChina and South Korea unveil breakthrough chips for real‑time brain interfaces and programmable photonicsChina's new brain‑speed chip enables sub‑10 ms neural processing, while South Korea's programmable photonic chip can dynamically control light speed, both promising advances for AI and medical technologies.
-
26 days ago
[TECHNOLOGY] 3 sourcesChina unveils record‑fast neuromorphic chip for AI and healthChinese scientists have built a 40 nm neuromorphic chip that mimics brain activity in real time, achieving up to 478× the speed of Nvidia A100 and promising breakthroughs in AI, brain surgery, early disease‑de
Sources
36kr.com · cnmo.com · diken.com.tr · finance.technews.tw · gazetedamga.com.tr · interestingengineering.com · kks.ch · risalehaber.com · scitechdaily.com · tr.sputniknews.com
This summary has been updated 1 time: see revision history