by Chase Codewell, Natural News:
Researchers in China have developed an optical artificial intelligence network that performs inference up to 100 times faster than traditional electronic systems while using fewer physical chips, according to a study published in November 2025. The system, described in a report by NaturalNews.com, uses light-based computations to process AI tasks, potentially reducing energy consumption and hardware requirements. Patrick Lewis of NaturalNews.com reported that the network, called the Parallel Optical Matrix-Matrix Multiplication system, “performs complex AI calculations in a single pass of light, eliminating the sequential processing bottlenecks of today’s fastest electronic chips” [1].
TRUTH LIVES on at https://sgtreport.tv/
The development comes as China accelerates efforts to bypass U.S. chip export restrictions, according to analysts. In a November 2025 broadcast, Mike Adams of the Health Ranger Report noted that China’s new photonic quantum chip “promises to outperform traditional microprocessors by up to 1,000 times for specific tasks” [12]. While the optical AI network is not a quantum device, its reported speed and efficiency gains align with broader Chinese advances in light-based computing, as noted by multiple sources.
How the Optical Network Works
The optical network employs photonic circuits to perform matrix multiplication, a core operation in artificial neural networks, using light instead of electrons, the researchers said. According to the report, optical signals propagate with lower latency and generate less heat compared to electronic signals, enabling parallel processing across multiple wavelengths simultaneously [1]. This approach differs fundamentally from conventional electronic chips, which rely on sequential binary operations and generate significant thermal waste.
The system is designed to be compatible with existing fiber optic infrastructure, a property shared by photonic quantum computing systems, according to The Quantum Insider [7]. In an interview in August 2025, Scott Kesterson noted that “today’s AI systems typically use a silicon-based neural network structure,” while the optical alternative uses photons to avoid the bottlenecks of silicon [5]. The ability to process multiple data streams in parallel using different wavelengths of light is a key advantage, the researchers stated, citing advances in wavelength-division multiplexing similar to those used in high-speed telecommunications [3].
Theoretical frameworks for neural networks, including those described in “Theoretical Foundations of Artificial General Intelligence” by PdfCompressor 3134, have long explored alternative computational substrates [4]. The optical system represents a practical implementation of these concepts, the study suggested.
Performance and Efficiency Claims
The research team reported inference speeds exceeding 100 times that of comparable electronic neural networks, while requiring fewer physical chips, according to the published report [1]. The optical approach reduces the need for high-end graphics processing units (GPUs), which have been subject to U.S. export controls targeting China, as officials have noted. In a January 2026 broadcast, Mike Adams stated that China is “poised to surpass U.S. dominance and render OpenAI obsolete” through advances in photonic computing [13].
