Q.ANT Unveils the Processor of the Future
1m, 2s
2
16:29, 25.11.2024
The Photonic Computing Revolution
German startup Q.ANT has announced the launch of its first commercial photonic processor, built on the unique LENA (Light Empowered Native Arithmetics) architecture. The processor uses light to perform complex nonlinear calculations, offering 30 times greater energy efficiency compared to traditional CMOS chips. With a standard PCI-Express interface, the new product is fully compatible with existing computing systems, ensuring seamless integration.
The Q.ANT processor is designed for high-performance tasks such as machine learning, neural network inference, computer vision, and complex physical simulations. According to the company’s CEO, photonic chips can reduce the energy consumption of a single GPT-4 query by a factor of 30 compared to current solutions.
Revolutionary Lithium Niobate Technology
A key feature of the chips is the use of a lithium niobate-based material, which Q.ANT has been developing since 2018. This technology replaces millions of transistors in traditional calculations, such as Fourier transforms, with a single optical element.
The first deliveries of the processors are scheduled for February 2025. Q.ANT offers the product as a ready-to-use Native Processing Server (NPS), which easily integrates into modern data centers and supercomputers via the proprietary Q.ANT Toolkit interface. This groundbreaking innovation promises to redefine the future of high-performance computing by enabling energy-efficient and ultra-fast technologies.
Was this article helpful to you?
Yes
No
VPS popular offers
sale
-20%
Linux
€ 8.8 /mo
Billed monthly
sale
-20%
Windows
€ 45.34 /mo
Billed monthly
-10%
€ 77.54 /mo
Billed annually
sale
-20%
Windows
€ 56 /mo
Billed monthly
sale
-20%
Linux
€ 6.1 /mo
Billed monthly
sale
-20%
Linux
€ 135.49 /mo
Billed monthly
sale
-20%
Linux
€ 8.3 /mo
Billed monthly
-15.3%
Windows
€ 54 /mo
Billed annually
sale
-20%
Windows
€ 54.49 /mo
Billed monthly
sale
-20%
Linux
€ 25.25 /mo
Billed monthly
