PCM+Crossbar approach
Electrical pulse modulation of phase-change material state with delay at nanosecond level.
Silicon-Photonics PCM Heterogeneous
Integration + Crossbar Photonic Matrix
Silicon-Photonics PCM Heterogeneous
Integration + Crossbar Photonic Matrix
Silicon-Photonics PCM Heterogeneous
Integration + Crossbar Photonic Matrix
Route Overview
Route Overview
Route Overview
Route Overview
Lightstandard adopts a technical approach based on heterogeneous integration of silicon photonics and phase-change materials, combined with a Crossbar photonic matrix computing architecture. By modulating the state of phase-change materials, we control the transmittance of light in waveguides containing phase-change materials. The incident light power and the phase-change material state can be mapped as multiplier and multiplicand respectively, and the power value of light after passing through the waveguide containing phase-change materials can be read as the product. Multiple product light beams converge into a single beam through coupling, and the power value of this beam represents the result after addition. Through the Crossbar (matrix cross) structure, multiple incident light beams can simultaneously perform multiplication and addition operations, thereby achieving matrix multiplication. Compared to other optical computing technical approaches, this approach enables in-memory computing with optical chips, offering advantages such as small unit size and low system power consumption.
Other approaches
Limited Size，Low Array Count，Complex Packaging
Limited Size，Low Array Count，Complex Packaging
Limited Size，Low Array Count，Complex Packaging
Unit size needs to reach 200-300um to achieve precise modulation, therefore optical computing chips of the same size have low integration density and low array quantity limits. Basic units require more than 10 electrodes, resulting in many wire bundles and high packaging difficulty.
PCM+Crossbar approach
Small Size, High Integration, Easy Packaging
Small Size, High Integration, Easy Packaging
Small Size, High Integration, Easy Packaging
PCM+Crossbar approach: Unit size is less than 1/10 of other approaches, therefore optical computing chips of the same size have high integration density with significant advantages in array quantity limits. The number of electrodes required for basic units is less than 1/5 of other approaches, resulting in fewer wire bundles and simpler packaging.
Zero Static Power Consumption, High Stability
Zero Static Power Consumption, High Stability
Zero Static Power Consumption, High Stability
PCM+Crossbar approach
Phase-change materials have non-volatility, requiring only one electrical drive to stably complete an AI computing task, resulting in zero static power consumption and no thermoelectric crosstalk issues.
Other approaches
Pure silicon materials require continuous heating to maintain state, resulting in not only high maintenance power consumption but also thermoelectric crosstalk that affects computing accuracy and stability.
High Modulation Speed, High Computing Efficiency
High Modulation Speed, High Computing Efficiency
High Modulation Speed, High Computing Efficiency
What industries are optical computing solutions best suited for?
What are the primary technical approaches in optical computing?
What are the advantages of optoelectronic fusion computing cards?
What emerging materials are utilized in optical computing chips?
What packaging architectures are available for optoelectronic fusion computing cards?
AI foundation models | Space-based computing | Artificial Intelligence Data Center | Embodied intelligence | Autonomous vehicles
Phase-change materials (PCM) | Sub-wavelength diffraction | Microring resonators (MRR) | Mach-Zehnder interferometers (MZI)
Low power consumption | Cost efficiency | High computational throughput | Superior stability
Glass substrates | Phase-change materials | Silicon photonics | Lithium niobate | III-V compound semiconductors
Co-packaged optics (CPO) | 3D packaging | 2.5D packaging | Monolithic integration
What industries are optical computing solutions best suited for?
What are the primary technical approaches in optical computing?
What packaging architectures are available for optoelectronic fusion computing cards?
AI foundation models | Space-based computing | Artificial Intelligence Data Center | Embodied intelligence | Autonomous vehicles
Phase-change materials (PCM) | Sub-wavelength diffraction | Microring resonators (MRR) | Mach-Zehnder interferometers (MZI)
Low power consumption | Cost efficiency | High computational throughput | Superior stability
Glass substrates | Phase-change materials | Silicon photonics | Lithium niobate | III-V compound semiconductors
Co-packaged optics (CPO) | 3D packaging | 2.5D packaging | Monolithic integration
Other approaches
Thermal modulation of silicon waveguide refractive index with delay at millisecond level.
