In a recent announcement, Cisco Systems (CSCO.O) introduced its latest AI networking chips that are specifically designed for AI supercomputers. These chips are expected to give tough competition to the existing offerings from Broadcom (AVGO.O) and Marvell Technology (MRVL.O).
Table of Contents
The Power of Cisco’s SiliconOne Series
Cisco has recently launched its SiliconOne series of chips as part of its efforts to enter the AI chip market. The chips are currently being tested by five out of the six major cloud providers. Top cloud service providers including Amazon Web Services, Microsoft Azure, and Google Cloud are reportedly being evaluated by Cisco, although the company did not reveal the names of the firms involved. Bofa Global Research has identified three cloud giants that hold a dominant position in the cloud computing market.
The significance of AI applications is growing rapidly.
Efficient communication between individual chips has been highlighted as critical with the growing popularity of AI applications, such as the well-known ChatGPT. Specialized chips, called graphics processing units (GPUs), are essential for powering the demanding workloads of AI supercomputers.
News: Cisco Unveils Latest Networking Solutions
Cisco, a well-known provider of networking equipment, offers a diverse array of solutions, including ethernet switches that enable effortless connectivity between different devices, such as laptops, computers, printers, routers, and servers, within a local area network.
“Enhanced Performance and Connectivity” – No further information was provided.
The latest generation of ethernet switches from Cisco, the G200 and G202, has been released with double the performance of their previous models. Advanced switches have been developed to meet the increasing demands of AI and machine learning workloads. These switches are capable of connecting up to 32,000 GPUs, offering an impressive capability.
“Fueling AI/ML Workloads” – No changes needed as it already sounds like a news headline.
Cisco fellow and former principal engineer, Rakesh Chopra, has announced that the G200 and G202 chips are the most powerful AI networking chips currently available on the market. Revolutionary chips are poised to transform AI and ML workloads with exceptional performance, promising an efficient and power-effective network.
Efficiency and Optimization
Innovative chips developed by Cisco have the potential to revolutionize AI and machine learning tasks. The chips can reduce the number of switches required by 40% and minimize latency, streamlining these processes. Enhanced efficiency contributes to both optimized performance and energy conservation, according to reports.
The Competition: Broadcom’s Jericho3-AI Chip
Broadcom’s Jericho3-AI chip made headlines in April upon its release. In the AI chip market, a formidable contender has emerged with a cutting-edge offering that can support the connection of up to 32,000 GPU chips.
Cisco has made a significant stride in the race for AI dominance with the release of its latest AI networking chips. This move is seen as a crucial step towards establishing the company’s foothold in the highly competitive AI arena. Cisco is targeting the increasing demands of AI supercomputers with their exceptional performance, connectivity, and efficiency. The aim is to accelerate advancements in the field of artificial intelligence.
