Photonic Computing
1. Processing & Core Computing
Silicon Photonics: The foundational technology of manufacturing photonic circuits on standard silicon substrates. It integrates optical components (lasers, modulators, detectors) directly onto silicon microchips to accelerate data transfer between electronic components.
Micro-ring modulators, used to switch and route optical signals in silicon photonics.
Photodiodes, used in regenerative latch designs to detect and convert light signals within the circuit.
Optical waveguides, used to carry signals across the chip.
Optical universal logic gates, which can be combined to implement latch behavior.
Photonic Processor: A microchip that uses light instead of electricity to perform mathematical calculations. By utilizing different wavelengths of light simultaneously, these processors can execute complex AI and matrix operations at extreme speeds with ultra-low power consumption.
Photonic Integrated Circuits (PICs): Chips that use light instead of electrical currents to route data, dramatically increasing bandwidth and reducing latency for AI clusters.
Photonic Latch (Memory): An optical equivalent of an electronic flip-flop circuit. It uses bistable optical states to lock and hold a bit of data (0 or 1) using light beams, serving as the ultra-fast, short-term memory (RAM) layer within a photonic computing architecture.
💾 2. Optical Data Storage
5D Optical Data Storage (Project Silica): The ultimate archival tier. It uses femtosecond lasers to create microscopic voxel structures inside quartz glass. The "5D" refers to the three spatial coordinates plus two optical dimensions (polarization and intensity). It is designed to preserve petabytes of data for billions of years without degrading.
CD / DVD / Blu-ray: The legacy consumer tiers of optical storage. They rely on surface-level pits and lands stamped onto polycarbonate discs, read by red or blue lasers. While revolutionary for their time, they suffer from limited capacity and material degradation (disc rot) over decades.
🌐 3. Networking & Communication
LiFi (Light Fidelity): A wireless communication technology that uses light (typically from LED bulbs) to transmit high-speed data through the air. It offers incredibly secure, interference-free, high-bandwidth local networking compared to traditional radio-frequency WiFi.
Fiber Optics: The backbone of the global internet. It guides light pulses through long, flexible strands of glass or plastic fibers over vast distances, enabling the high-speed global telecommunications infrastructure we rely on today.
Optical Transceivers & Engines: High-speed modules (reaching 1.6 Terabits per second and beyond) that convert electrical signals to optical signals for data transmission.
Co-Packaged Optics (CPO): Placing optical engines directly next to processors like GPUs or switches to eliminate signal loss over short copper distances.
📷 4. Optical Sensing
Camera: The primary input mechanism for capturing optical information from the physical world. In the context of a purely photonic system, modern advancements are moving toward co-designing camera sensors directly with photonic processors to decode, filter, and process images instantly before the data ever reaches a traditional electronic CPU.
Leading photonic computing companies include Lightmatter, PsiQuantum, and Xanadu. These firms are advancing light-based processors for AI, quantum applications, and high-speed data processing as of 2026.
Lightmatter leads in photonic AI chips deployed in data centers, offering massive speedups for tensor operations. PsiQuantum develops fault-tolerant quantum systems using silicon photonics, backed by billions in funding and partnerships like GlobalFoundries.
Xanadu advances squeezed-light photonic quantum processors with cloud access and error correction milestones.
Feynman Platform 2028
At GTC 2026, NVIDIA provided the first deep technical dive into the Feynman architecture due to be launched in 2028. Following the Blackwell and Rubin generations, Feynman is designed to be the backbone of "Gigawatt-scale" AI factories. The architecture shifts from simply connecting GPUs to a "system-of-systems" approach, utilizing advanced 3D stacking and optical interconnects to overcome the physical limits of copper.
In the Feynman platform, scale-up transitions from copper-based backplanes to native CPO. Feynman GPUs are designed to integrate silicon photonics directly onto the GPU package. NVIDIA is working on developing standalone I/O optical chips as well as integrated chips where the laser is integrated into the I/O chip. However, initial designs almost certainly will use remote rather than integrated lasers for reliability and serviceability reasons.
NVIDIA is also expected to use the Optical Compute Interconnect (OCI) standard to attach CPO modules to its GPUs. Launched at OFC26 last month, the OCI Multi-Source Agreement (MSA) is one of several optical interconnect MSAs launched recently. NVIDIA has positioned itself as a founding member of the OCI MSA to ensure that all the optical components – lasers, fibers and silicon photonic connectors – become part of a standardized multi-vendor ecosystem.
Psi Quantum - Photonic Quantum Computer
A Photonic Integrated Circuit (PIC) is a microchip that uses light (photons) instead of electricity (electrons) to generate, transport, process, and measure information.
PICs
Waveguides: Act like tiny optical wires, guiding light across the chip using total internal reflection.
Active Components: Include lasers for light sources, modulators to encode data onto the light beam, and photodetectors to convert light back into electrical signals.
Passive Components: Include splitters, filters, and resonators that route and shape the optical signals without needing external power.
Photonic Latch
Stores one bit of information in two stable states, like a flip-flop or latch in digital electronics. Uses silicon photonic micro-ring modulators and universal optical logic gates to implement the memory behavior. Supports optical set/reset outputs and complementary outputs, which makes it useful for optical computing and communications. It is designed to be fast and scalable, with picosecond-scale response reported in coverage of the prototype. It is compatible with standard silicon photonic fabrication, which improves the chance of practical manufacturing.
For processor-oriented photonic systems, the main available building blocks are:
Energy Efficiency: Photonic hardware can process matrix multiplications—the core operation of AI models—using significantly less power than electrical chips, potentially reducing energy consumption by over (90%) for certain operations.
Speed and Latency: Photonic processors operate at the speed of light, overcoming the limitations of copper wires, which face signal loss and heat issues at scale.
Scalability: Photonics offers immense bandwidth, essential for the next generation of large-scale AI factories, with research showing over (96%) accuracy in AI tasks.
Main technology
Compute
Photonic processors, interferometer meshes, optical accelerators
On-chip I/O
Silicon photonics, waveguides, microrings, photodiodes
Memory/state
Optical latches, resonators, delay lines
Network
Fiber, transceivers, CPO, LiFi
Sensing
Cameras, optical sensors, optical preprocessing
Hard Drive
5D glass storage / Project Silica Permanently records data in glass