The launch positions Cisco more directly against rivals Broadcom and Nvidia as companies race to capture a share of what industry executives estimate to be a $600 billion wave of spending on AI infrastructure.
At the center of Cisco’s push is the Silicon One G300 switch chip, which the company expects to bring to market in the second half of the year. The chip is designed to help the processors that train and run AI models communicate efficiently across data centers containing tens or even hundreds of thousands of interconnected links.
Manufactured using Taiwan Semiconductor Manufacturing Co’s advanced 3-nanometer process, the G300 incorporates new “shock absorber” features intended to prevent network slowdowns during sudden surges in data traffic. Those congestion issues are a growing challenge as AI workloads place extreme and unpredictable demands on data center networks.
Cisco says the chip can accelerate certain AI computing tasks by as much as 28%, partly by automatically rerouting data around network bottlenecks within microseconds. According to Martin Lund, executive vice president of Cisco’s common hardware group, the focus is on maintaining overall system efficiency rather than just raw speed.
“When you have tens of thousands or hundreds of thousands of connections, disruptions happen quite regularly,” Lund said in an interview. “We focus on the total end-to-end efficiency of the network.”
Networking has emerged as a critical battleground in the AI ecosystem, alongside graphics processors and custom accelerators. Nvidia highlighted the importance of this segment last month when it introduced new AI systems that include a dedicated networking chip competing directly with Cisco’s products. Broadcom is also targeting the same space with its Tomahawk line of switching chips.
As AI models grow larger and more complex, the ability to move data quickly and reliably between chips is becoming just as important as the computing power of the chips themselves. Cisco’s latest launch underscores how traditional networking companies are repositioning themselves to stay relevant in an AI-driven data center market that is evolving at breakneck speed.