Best U.S. Networking Stocks to Own Into Q3 Earnings: Deutsche Bank Favors Cisco and Arista
Artificial intelligence is creating a new investment opportunity beyond GPUs and data-center servers: networking. As AI clusters become larger and more interconnected, the amount of data moving between computing systems is rising rapidly, increasing demand for faster switches, optical components and high-speed connectivity.
Against this backdrop, Deutsche Bank analyst Gianmarco Conti has highlighted Cisco Systems (NASDAQ: CSCO) and Arista Networks (NYSE: ANET) as its preferred U.S. networking stocks heading into the third-quarter earnings season. Both stocks have a Buy rating from the bank, although Deutsche Bank also sees an emerging technology risk that investors should not ignore.
The key question for investors is whether AI-driven networking demand can continue to outweigh the potential disruption from optical circuit switching.
Why networking has become an AI investment theme
AI infrastructure is increasingly becoming a networking problem as much as a computing problem.
Training and running large AI models requires thousands of GPUs and other processors to communicate rapidly. That creates demand for higher-speed connections between servers, racks and data centers.
The opportunity is particularly important as AI infrastructure expands from traditional scale-out architectures, where computing resources are added across larger clusters, toward increasingly sophisticated scale-up and rack-scale systems.
Deutsche Bank expects rising port speeds and increasing connectivity requirements to support networking demand even as the underlying architecture evolves. The bank also believes that AI inference — when trained models are used to generate answers or predictions — remains difficult to handle without packet switching because workloads can be less predictable than large training operations.
That distinction matters for networking companies because it suggests that AI growth may not automatically translate into a replacement of conventional switching technology.
Cisco: The diversified way to play AI networking
Cisco is the more established name of the two.
The company has been benefiting from stronger demand across its networking portfolio while also building a larger position in AI infrastructure. Cisco's fiscal fourth-quarter 2026 results showed the scale of the opportunity: revenue reached $17.3 billion, up 18% year over year, while networking product revenue increased 28%.
More importantly for the AI thesis, Cisco said it received $4 billion of AI-infrastructure orders in the fourth quarter, taking fiscal 2026 orders to $9.3 billion. The company expects approximately $7.5 billion of AI-infrastructure revenue in fiscal 2027.
Cisco therefore offers investors exposure to AI networking without relying entirely on one narrow market.
Its business spans networking, security, observability, collaboration and other enterprise technologies. That diversification can be valuable if spending in one part of the technology market slows.
Deutsche Bank's argument is that Cisco's broad portfolio can continue to benefit from the need for higher-speed networking as AI clusters expand. The bank also expects growth in larger-scale connectivity to help offset potential pressure from changes in traditional data-center architectures.
Arista Networks: More direct exposure to the AI networking cycle
Arista is the more focused AI-networking opportunity.
The company specializes in high-performance networking for cloud, data-center, AI and routing environments. Its products are particularly relevant to large-scale computing clusters where extremely fast communication between processors is essential.
Arista's Q2 2026 results demonstrated the strength of that demand. The company reported $3.04 billion in quarterly revenue, up about 38% year over year, while adjusted earnings per share came in at $1.02.
The company has also continued expanding its AI-networking technology. In June, Arista introduced a 1.6-terabit networking platform designed for rack-scale AI infrastructure.
Arista has scheduled its Q3 2026 results for November 3, 2026, after the U.S. market closes. Management will discuss the results during a conference call later that day.
That makes the upcoming report an important test of whether AI-networking demand remains as strong as investors currently expect.
The risk Deutsche Bank is watching
The bullish networking thesis is not without a potential technological challenge.
Deutsche Bank has identified optical circuit switching (OCS) as a possible threat to traditional networking architectures, particularly at the spine layer. OCS uses optical connections to establish direct communication paths and could potentially reduce the need for some conventional packet-switching infrastructure in certain AI workloads.
For networking investors, this is an important distinction.
A change in architecture does not necessarily mean that networking demand disappears. Instead, the mix of equipment required could change.
Deutsche Bank's view is that the broader opportunity remains intact because AI inference still requires flexible packet switching. The bank also expects growth in "scale-across" infrastructure — connecting larger pools of computing resources — to provide another source of demand.
This makes the technology transition something investors should monitor rather than automatically treat as a bearish signal.
What investors should watch in Q3 earnings
For Cisco and Arista, the headline revenue and earnings numbers will be important, but the commentary around AI infrastructure could matter even more.
Investors should pay particular attention to:
AI-related orders and revenue: Are hyperscaler and AI-infrastructure customers continuing to increase spending?
Networking growth: Is high-speed networking growing faster than the broader business?
Port speeds: How quickly are customers moving toward higher-speed networking platforms?
Hyperscaler spending: Are major cloud providers maintaining their infrastructure investment plans?
AI inference: Is inference creating incremental demand for packet-switching technology?
OCS adoption: Are customers beginning to shift meaningful workloads toward optical circuit switching?
Margins: Can networking companies maintain profitability while investing heavily in next-generation infrastructure?
Cisco's recent results already provide evidence of strong networking momentum. The company said networking product orders increased 40% year over year in fiscal Q4, its eighth consecutive quarter of double-digit growth in networking product orders.
Arista's upcoming November report will provide another important data point for the AI-networking cycle.
Cisco vs. Arista: Different ways to approach the same trend
The two stocks offer different risk profiles.
Cisco is the more diversified choice. Its established enterprise customer base, networking franchise and expanding AI infrastructure business give investors exposure to the AI networking theme while reducing dependence on a single growth driver.
Arista, meanwhile, offers more direct exposure to high-performance cloud and AI networking. That creates greater sensitivity to AI infrastructure spending and potentially greater upside if the networking cycle remains strong, but it also leaves the stock more exposed to changes in hyperscaler capital expenditure and networking architecture.
Neither approach is automatically superior. The right choice depends on whether an investor prioritizes diversification or more concentrated exposure to the AI networking opportunity.
The bottom line for investors
Deutsche Bank's preference for Cisco and Arista Networks reflects a broader shift in the AI investment story: as computing capacity expands, moving data between those systems is becoming increasingly important.
The immediate catalyst is the upcoming earnings season. Cisco's recent results already point to strong networking and AI-infrastructure demand, while Arista's November 3 report will offer another important read on the strength of the cycle.
The biggest issue to watch is not simply whether networking demand grows, but which networking architecture wins as AI infrastructure evolves. Packet switching, higher-speed ports and emerging optical technologies could all play a role.
For investors, Cisco and Arista therefore remain compelling names to monitor, but the next round of earnings should be judged on demand quality, AI spending visibility, margins and evidence of any architectural shift toward optical circuit switching.
Follow the blog for more updates and analysis on U.S. stocks, AI infrastructure, technology companies and global markets.
This article is for informational and educational purposes only and should not be considered investment advice.