Cisco Ramps Its Partnership with NVIDIA—and Both Stocks Are Soaring

Cisco's partnership with NVIDIA is proving to be a winner—and it's expanding.

With announcements from NVIDIA’s GTC conference in Washington D.C., today, NVIDIA and Cisco firmed up their partnership with new integrated wares targeted at the rich and fast-developing AI and datacenter space. As part of the announcement, NVIDIA will enable Cisco Silicon One coupled with NVIDIA SuperNICs to become part of the NVIDIA Spectrum-X Ethernet networking platform. Cisco said that for now, it will be the only partner silicon included in NVIDIA Spectrum-X.

Cisco will also build systems combining NVIDIA Spectrum silicon with Cisco operating system software, allowing customers to use a combined Cisco networking and NVIDIA platform in the datacenter. This will include the Cisco N9100 switch.

Both NVIDIA and Cisco shares are trading at yearly highs, with NVIDIA shares today hitting an all-time high. Cisco shares are up 22% year to date. As two of the giants in the cloud and enterprise infrastructure markets, this partnership is a show of strength, combining NVIDIA's lead in AI compute and Cisco's dominance in enterprise networking. By partnering with Cisco, NVIDIA gets access to its networking silicon, operating systems, and powerful enterprise channel. By partnering with NVIDIA, Cisco can integrate its networking gear into AI clusters for both cloud and enterprise markets.

Source: Google

Targeting Both Enterprise and Cloud

As part of the announcements, a number of products were announced for different markets. The breadth of the offerings—targeting markets as diverse as enterprise verticals, neoclouds, and telecom—shows that both companies are feeling bullish about the deal partnership they announced last February. Here at Futuriom, we viewed that and other actions by Cisco as a turning point for the company. The idea of combining Cisco's enterprise networking acumen with NVIDIA's AI compute and software expertise provides an "easy button" for AI adoption. The two companies signaled an expansion of the deal in June.

The Cisco N9100 series switches will now offer a choice of Cisco NX-OS or SONiC operating systems, which the company says offers flexibility for growth markets from neoclouds and sovereign cloud customers to enterprise. Cisco’s Nexus porfolio will now ship on a number of silicon platforms, including Cisco’s own Silicon One, as well as NVIDIA’s Spectrum-X Ethernet switch silicon. The N9100 scales to 51.2 Tbps of bandwidth.

Will Eatherton, Senior Vice President of Cisco Networking Engineering, told Futuriom that the new product portfolio is about having the scale and flexibility for a wide range of markets, from enterprise to neoclouds.

"With the Cisco N9100 Series, we now introduce to the market a NVIDIA Cloud Partner(NCP)-compliant reference architecture," wrote Eatherton in his blog today. "This development is particularly significant for neocloud and sovereign cloud customers building data centers with capacities ranging from thousands to potentially hundreds of thousands of GPUs, as it allows them to diversify their supply chains effectively."

In addition to neoclouds and sovereign buildouts, Cisco and NVIDIA are also targeting private enterprise deployments, for businesses that want to build their own AI infrastructure instead of relying on the cloud. It's all part of an explosion of interest in many new areas of AI networking, with the proof being recent successes by other networking companies, including Arista and Nokia.

For enterprise customers, Cisco has announced the Secure AI Factory with NVIDIA with security and observability integrations. Cisco and NVIDIA are also looking to move deeper into the telecom market, with what they are calling the first AI-native wireless stack for 6G. The product portfolio also shows an emphasis on Ethernet, with the customer able to choose from a number of flavors and operating systems

The Secure AI Factory checks the boxes for the enterprise market, with an emphasis on security. It includes integration with NVIDIA AI Enterprise software to deliver robust cybersecurity for AI applications using NVIDIA NeMo Guardrails. NVIDIA AI Enterprise is one of the top six leading platforms we have identified in our AI Enterprise Index of enterprise case studies. The product can also be combined with Cisco AI Defense for on-premises deployment enabling security and AI teams to protect AI models and applications.

Building for the Rack-Scale World

The expanded deal between Cisco and NVIDIA also reinforces a strong new trend in the industry. The major hardware and networking OEMs are ramping up integrations with a wide variety of compute, networking, and storage solutions to build powerful, turnkey solutions for markets such as private enterprise AI and neocloud.

One of the challenges for many industries in adopting AI is the complexity and learning curve. The large OEMs believe they can help solve that with ready-made, "rack-scale" solutions.

For example, some boxes have been optimized for telecom and edge/cloud convergence strategies. In the neocloud world, partners are developing rack-scale boxes with the highest speeds and feeds. And in the enterprise, the focus on is on flexibility and economy with high-level security for apps and data.

This is a powerful trend that Futuriom analysts have seen accelerate in recent weeks, with massive rack-scale devices unveiled at OCP and GTC DC. For example, Dell last week announced updates to the the Dell AI Data Platform, the subset of the Dell AI Factory that deals with storage and data handling.

Dell’s announcements also show how the AI Data Platform is in competition with storage vendors—or companies perceived as storage vendors—that are extending into data management. Specifically, Dell wants to call out its differences from VAST Data and Pure Storage.

Interestingly, the NVIDIA and Cisco deal comes at a time of accelerating deals for networking for AI. On the same day, NVIDIA also announced a deal to invest in Nokia, which recently reported outstanding results it attributed to growth in networking for AI datacenters.

(Futuriom Take: This eye-opening announcement is a demonstration of power among the leading AI compute and enterprise networking players. It also reinforces the need for OEMs to bring integrated, rack-scale compute and networking solutions to market to enable AI adoption.)