Qant Develops Photonic Chips to Reduce AI Data Center Energy Consumption by 90%
Stuttgart-based startup Qant is developing photonic chips that could reduce energy consumption in AI data centers by up to 90%. Utilizing light instead of electricity, these chips perform complex tasks more efficiently than traditional processors, with clock speeds reaching 30 GHz. Qant is currently producing 50,000 to 60,000 chips annually and aims to expand into commercial markets while collaborating with research institutions on algorithm development.
Theia Market Signal Identification - AI Assisted
Stuttgart-based startup Qant is developing photonic chips that could reduce the energy consumption of AI data centers by up to 90%. CEO Michael Förtsch emphasizes that traditional chips are inefficient for complex tasks, leading to excessive energy use.
Qant's photonic processors, utilizing light instead of electricity, can perform complex functions with significantly less overhead. The company is currently producing 50,000 to 60,000 chips annually, with a portion already delivered to research centers.
These chips are designed to work with existing silicon platforms but utilize lithium niobate thin films for precise control. Qant aims to expand into commercial markets while collaborating with research institutions to develop suitable algorithms. The second-generation processors show significant advancements, reaching clock speeds of up to 30 GHz and matching the precision of digital circuits.
Comments