Photonics startup Neurophos, incubated at Duke University, raised $110 million in Series A funding led by Gates Frontier to develop its "super-surface modulator" optical processing unit (OPU). The company claims that its chip uses optical properties for artificial intelligence matrix calculations, operating at a frequency of up to 56GHz, with a processing speed of 235 million billion operations per second (235 Peta Operations per Second), and a power consumption of only 675 watts. It significantly outperforms NVIDIA's B200 GPU in both AI inference speed and energy efficiency. Neurophos expects its first chips to be available by mid-2028 and states that these chips can be manufactured using standard silicon wafer foundry processes.
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
Photonics startup Neurophos, incubated at Duke University, raised $110 million in Series A funding led by Gates Frontier to develop its "super-surface modulator" optical processing unit (OPU). The company claims that its chip uses optical properties for artificial intelligence matrix calculations, operating at a frequency of up to 56GHz, with a processing speed of 235 million billion operations per second (235 Peta Operations per Second), and a power consumption of only 675 watts. It significantly outperforms NVIDIA's B200 GPU in both AI inference speed and energy efficiency. Neurophos expects its first chips to be available by mid-2028 and states that these chips can be manufactured using standard silicon wafer foundry processes.