It's been the year of AI, and expectations are high as enterprise network vendors begin to deliver on the promise of AI networking. Industry players have long been touting the benefits of AI-based operations (AIOps) to help resource-constrained network teams troubleshoot and remediate performance issues, and modern network services such as SD-WAN, 5G and SASE rely on AI to automate some network provisioning and security tasks. Now the release of Microsoft-funded ChatGPT has elevated AI to a C-level imperative, and AI innovation is heating up.
Everyone is jumping on the bandwagon. Nvidia is cranking out GPUs for AI workloads at a record pace, and the hyperscalers are gobbling them up to enhance their AI offerings. Enterprises are racing to embed AI functionality throughout their on-prem and cloud infrastructures. For enterprise network vendors, there's an opportunity to deliver products and services with an extra helping of generative AI smarts.
On a more fundamental level, those GPUs need to move incredible amounts of data, which creates demand for high-performance interconnects. InfiniBand is commonly used for high-performance computing, but Cisco, Arista, Broadcom, HPE and others have created the Ultra Ethernet Consortium with the goal of making Ethernet the industry standard for AI networking.
The 10 power players on this list are able to deliver highly mature offerings in areas of LAN and WAN infrastructure and services, have a strong security play, and are positioning themselves to take full advantage of the AI boom.
Editor's note: Our list is ordered, with input from industry watchers, to reflect the companies that are making the biggest power moves and the broadest impact on the network industry. We’ve evaluated these companies not by size but by their influence in the market and their forward momentum. In other words, there is some subjectivity associated with this list.
1. Cisco: Plunks down cash for Splunk
Why they're here: When you're the dominant vendor in your market seemingly forever, there might be a tendency to get complacent. Not Cisco under CEO Chuck Robbins. Cisco shook up the networking world with the announcement in September that it plans to buy SIEM stalwart Splunk. Robbins explained that integrating Splunk's sophisticated data analytics capabilities will bolster Cisco's AI push. The company reported that it has already taken $500 million in orders for AI Ethernet fabrics, mostly from hyperscalers, and sees the potential for much more. Overall, Cisco continues to run on all cylinders - Ethernet switch revenues increased 55.3% year over year in 2Q23, giving the company a market share of 47.2%. Most recently, Cisco said its backlog is down, and in its fiscal Q1 2024 report, Robbins said the company achieved the strongest Q1 results in its history in both revenue ($14.7 billion) and profitability ($4.5 billion). "After customers implement large amounts of recently shipped product, we expect to see product order growth rates accelerate in the second half of the year," Robbins stated.
Power moves: Aside from the Splunk splash, Cisco has acquired four other companies focused on security: Armorblox, a threat detection platform; Oort, which does identity management; and Valtix and Lightspin, both in cloud security.
By the numbers: $28 billion. The purchase price for Splunk.
Outlook: Cisco recently released a new security service edge (SSE) offering, a new service designed to protect multicloud workloads and a new high-end firewall. But the focus is on AI. Robbins said: "The acceleration of AI will fundamentally change our world and create new growth drivers for us. Cisco’s ASIC design and scalable fabric for AI position us very well to build out the infrastructure that hyperscalers and others need to build AI ML clusters. We expect Ethernet will lead in connecting AI workloads over the next five years."
2. Broadcom: Seals the deal for VMware
Why they're here: The $61 billion purchase of VMware, finalized on Nov. 22, catapults Broadcom into the top echelon of networking power players. The company has cobbled together an impressive lineup of products and services, including semiconductors, storage infrastructure, networking (Brocade), network management (CA) and security (Symantec.) The goal of the VMware deal is to provide enterprises with a way to extend on-prem virtualized workloads to multi-cloud environments.
Power moves: Broadcom bought ConnectALL to expand its value stream management (VSM) capabilities. Research firm IDC reports that VSM has become a high priority investment area for enterprises because it helps companies assess the value generated by software development initiatives.
By the numbers: $2 billion. Seeking to reassure VMware customers anxious about what will happen after the acquisition, Broadcom CEO Hock Tan pledged that the company will spend $2 billion a year to accelerate research and development for VMware products and services.
Outlook: Tan has outlined a clear strategy for Broadcom going forward. He said: "We will invest in extending VMware's software stack to run and manage workloads across private and public clouds, which means any enterprise can run application workloads easily, securely and seamlessly on-prem, or in any cloud platform they prefer." The increased investment will also be focused on building VMware's professional services. Gaining acquisition approval from regulators in China was a significant hurdle for Broadcom to clear; now it needs to absorb VMware without losing customers to competitors that are eagerly offering VMware migration assistance.
3. Arista: Hitting on Ullal cylinders
Why they're here: Arista is a powerhouse in high-performance networking aimed at the most demanding data center and cloud environments. Gartner said Arista is a visionary in wired and wireless infrastructure with its leaf and spine switches, wireless access points, and CloudVision management platform, which provides network automation, integrated security and AI/ML capabilities. Ethernet switch revenue skyrocketed 42.6% year over year, according to IDC's latest tracker, giving the company a 10.4% market share. Third-quarter revenue hit $1.5 billion, up 28% from the third quarter of 2022.
Power moves: Arista recently announced its first foray into the WAN with a product called WAN Routing Systems that is expected to compete in the SD-WAN market. Specific components include hardware routers, virtual and cloud routers, and a WAN management service. Arista has also partnered with Equinix to deploy the WAN routing system in Equinix data centers.
By the numbers: 75 million. Number of cumulative cloud networking ports sold.
Outlook: Similar to Cisco, Arista envisions Ethernet becoming the network standard for AI. The company is one of the founders of the Ultra Ethernet Alliance. Arista President and CEO Jayshree Ullal said: "Generative AI Applications are pushing the envelope of networking scale akin to using all highway lanes simultaneously and efficiently. Once again, Ethernet will ultimately emerge as the winner in networking for AI. Together with IP, Ethernet will drive numerous use cases for AI training and inference. Scalable and efficient mechanisms implemented with packet spraying, flexible ordering and modern congestion control algorithms will be infused into AI-based Ethernet and IP networks. Welcome to the new decade of AI networking."
4. Palo Alto Networks: Seeking SOC transformation
Why they're here: Palo Alto Networks, a longtime power player in next-generation firewalls, has built out an integrated on-prem and cloud security platform. Palo Also is a leader in SD-WAN, SSE and single-vendor SASE, according to Gartner. And the company has an AI play, its extended security intelligence and automation management offering called XSIAM, which is designed to replace traditional SIEM and related point products with a more integrated, automated solution. Palo Alto is also focused on extending security to the software development process with the goal of building security into the full application lifecycle.
Power moves: Palo Alto bought Cider Security, a leader in application security and supply chain security.
By the numbers: $1.9 billion. Palo Alto reported fiscal first-quarter 2024 revenue grew 20% year over year to hit $1.9 billion.
Outlook: CNBC's Jim Cramer recently described Palo Alto as "just kind of unstoppable." Fiscal year 2023 revenue grew 25% to $6.9 billion, and the company is predicting revenue growth in the 16-19% range going forward. “We finished off the year with strong execution and the changing environment drove more customers towards platformization,” said Nikesh Arora, chairman and CEO of Palo Alto Networks. “Our strategy is resonating with a growing number of our customers, driving continued consolidation, to deliver superior security outcomes. We were delighted with the reception in the market for our AI-based security automation platform, XSIAM.” Palo Alto doesn't just want to offer incremental improvements to enterprise security; the company talks about transformational change through offerings like Cortex XSIAM, which is designed to fundamentally change how security operations centers (SOC) operate. The goal is to embed automation and analytics to reduce manual tasks and create an AI-driven, mostly autonomous SOC.
5. HPE (Aruba): Leaning heavily into AI
Why they're here: AI is at the forefront of HPE's efforts, whether that's building AIOps functionality into Aruba Central, its wired and wireless network management platform, or adding large language model (LLM) offerings to GreenLake, HPE's consumption-based service for edge, on-prem and hybrid cloud computing. HPE (Aruba) is a leader in Gartner's Magic Quadrant for SD-WAN and also a leader in Gartner's Magic Quadrant for wired and wireless infrastructure, with its CX switches and Wi-Fi 6/6e access points.
Power moves: HPE bought a slew of companies, including OpsRamp, an IT operations management company; Athonet, a leader in private 5G cellular networking; Pachyderm, whose software automates machine learning pipelines for large-scale AI applications; TidalScale, whose technology creates software-defined servers; and Axis Security, which provides an SSE platform.
By the numbers: $1 billion. The annualized run rate for GreenLake.
Outlook: HPE continues to chart its own path when it comes to generative AI, giving enterprises an alternative to building out AI functionality themselves, or going with a hyperscaler. The company recently announced HPE GreenLake for Large Language Models, which enables enterprises to train, tune, and deploy large-scale AI on a supercomputing platform that combines HPE's AI software and hardware. The platform includes an LLM called Luminous from German startup Aleph Alpha running on HPE Cray supercomputers in colocation facilities. HPE plans to launch AI-based applications in healthcare and life sciences, financial services, manufacturing and transportation. Said CEO Antonio Neri: "We have reached a generational market shift in AI that will be as transformational as the web, mobile and cloud. HPE is making AI, once the domain of well-funded government labs and the global cloud giants, accessible to all."
6. Fortinet: AI-powered networking and security
Why they're here: Sitting at the convergence of networking and security, Fortinet has generally followed the approach of building out its own product line, so the "Forti" brand of products and services is fully integrated from the jump. Gartner calls Fortinet a visionary in wired and wireless LANs. "Its FortiAP and FortiSwitch products are broadly focused on tight integration with network security capabilities leveraging its FortiGate security appliances and FortiCloud and FortiLAN cloud-based management platforms." Fortinet is also leader in Gartner's Magic Quadrant for SD-WAN. Most recently, the company unveiled Fortinet Advisor, an AI-based assistant aimed at helping security operations teams make more informed decisions, respond to threats faster, and simplify routine and complex tasks.
Power moves: Teamed up with Google to expand its SASE points of presence (PoP) across Google Cloud's global network edge locations.
By the numbers: 31,000. The estimated number of Fortinet's SD-WAN enterprise customers.
Outlook: In its latest reporting period, Fortinet's quarterly revenue increased 16% year-over-year to $1.3 billion, and its service revenue, in particular, rose 28%. The company continues to bolster its product portfolio. It recently announced that FortiSASE, the company's secure access service edge offering, is now integrated with its wireless LAN portfolio with the goal of providing more options for remote workers and distributed edges to connect securely. John Maddison, chief marketing officer and executive vice president of product strategy, said: "We are the only vendor offering fully converged wired and wireless networking and AI-powered security through a single platform, providing visibility and security from the moment a user or device connects to the network.
7. Juniper: Mist opportunities
Why they're here: Juniper has been ahead of the curve when it comes to AI. Juniper bought Mist Systems in 2019 and has been integrating its AI capabilities throughout the Juniper portfolio of routers, switches, wireless access points, firewalls, networking software and cloud-based management systems. Gartner gives Juniper high marks for AI-driven automation, particularly its Marvis virtual network assistant. Juniper said that its AI-driven enterprise revenue is growing 17% year over the year, led by Mist AI, which grew more than 60% year over year. However, Juniper is also facing headwinds; its service provider and cloud revenues have been lagging. And according to IDC, Juniper's router revenue only increased by 2.5% year over year in the second quarter, giving it a 10.3% market share. And while its Ethernet switch revenue grew 35.2% year over year, that didn't move the needle much - Juniper's market share remains below 3%.
Power moves: Juniper announced a $59 million restructuring plan aimed at putting more emphasis on its enterprise networking business, which has been growing faster than service provider or cloud segments.
By the numbers: $433 million. The first quarter of 2023 marked the first time in Juniper's history that its enterprise networking business was the largest of its three core divisions, growing at 18% to reach $433 million.
Outlook: Juniper continues to lead the way in AI-based networking. It recently announced the integration of the ChatGPT AI-based LLM with the Marvis VNA. It has added AI-based predictive threat support to its new family of firewalls. Juniper CEO Rami Rahim said: "AI networking represents a once-in-a-generation inflection point that will present us with complex technical challenges for years to come. And I believe we have the pieces at Juniper to enable this future."
8. Extreme Networks: Extremely focused on integration
Why they're here: Extreme's One Network approach combines wired and wireless networking, plus management software for on-prem and cloud. Gartner said Extreme's strengths are network fabric automation, digital twin capabilities and multivendor integration with Extreme Cloud, its management platform. Extreme is also a leader in AIOps capabilities, which brings the power of AI to help optimize IT operations. The company's fiscal-year 2023 earnings report reflects the strong momentum that Extreme has built: revenue was up 18% for the full year to reach $1.3 billion. In Q1 2024, revenue hit $353 million, up 19% year-over-year.
Power moves: Extreme CEO Ed Meyercord hasn't been shy about publicly calling out Cisco. He boldly said that Extreme offers a fully integrated alternative to Cisco's multiple product lines.
By the numbers: 650. Number of new employees Extreme hired over the past year, many from competing network vendors.
Outlook: Rosenblatt Securities analyst Mike Genovese said: "We think there is fatigue with Cisco’s products in the market, and Extreme’s solutions are strong enough to compare favorably to Juniper’s." Meyercord is also bullish on the future: "We outperformed our original top-line outlook for fiscal 2023, and based on industry analysts, estimates outgrew the market by two times. This, combined with the increase in volume of larger deals and new logos, is a clear indication that we're taking share from our largest competitors." He adds, "Our AIOps solutions are getting traction with customers as they look for new ways to leverage the network to drive better business outcomes."
9. Dell: Bringing AI to on-prem environments
Why they're here: With its powerful portfolio of high-performance computers, storage systems and networking, Dell is perfectly positioned to provide enterprises with a full-stack, on-prem, AI solution. Toward that end, Dell has teamed up with Nvidia to launch Project Helix, designed to help enterprises build and manage generative AI models on-premises. The offering combines Nvidia GPUs, LLM, software and networking with Dell PowerEdge servers and object storage systems. To make generative AI even more accessible to enterprises, Dell is offering preconfigured and pre-tested designs tailored to specific use cases. Project Helix meshes nicely with Dell's Apex as-a-service consumption pricing model, so enterprises can take advantage of generative AI without the upfront capital expenditures.
Power moves: Bought Moogsoft for its AIOps platform that helps companies troubleshoot technical issues in their infrastructure.
By the numbers: $100 million. The amount Dell paid to acquire Cloudify Ltd, whose technology will help Dell enhance its edge computing offerings.
Outlook: Dell positions itself an attractive option for enterprises that want to reduce the complexity associated with a multicloud environment. Dell is teaming up with Microsoft on Dell Apex Cloud Platform for Microsoft Azure, an offering that provides a jointly engineered on-prem appliance. Said Dell's Caitlin Gordon: "The whole magic of this is that we have integrated the full stack from the firmware all the way up to the Microsoft software, and that's fully automated." Dell is also planning to deliver Apex Cloud Platforms for VMware and Red Hat OpenShift.
10. Nvidia: Data center dominance
Why they're here: Gaming chip maker Nvidia bet on AI and the company's foresight, planning and execution have paid huge dividends. The company built out a full-stack AI offering that includes its own GPUs, LLM, CUDA AI software, plus networking and switch technology from its acquisition of Mellanox. Hyperscalers are snapping up Nvidia GPUs for their own data centers and Nvidia has teamed up with Dell to offer enterprises an on-prem AI platform.
Power moves: Announced its DGX cloud service, running now on the Oracle cloud and expected to be available soon on Microsoft Azure and Google Cloud. DGX Cloud is a hardware and software package that enables enterprises to create generative AI models using Nvidia technology inside the hyperscaler's environment,
By the numbers: $14.5 billion. Nvidia's data center revenue hit $14.5 billion in its latest quarter. Data center revenue is now 80% of the company's total revenue.
Outlook: Nvidia continues to look for ways to spread its AI technology to new markets. It is working with ServiceNow and Accenture on a platform to enable enterprises to build custom AI apps. It has an AI platform for digital twin and robotics technologies. And Nvidia is building specialized chips for the automotive industry. Alexander Harrowell, principal analyst at Omdia, said: "There are plenty of companies that have a powerful neural-network accelerator chip, but there is only one that has Nvidia's software ecosystem." Forrester analyst Glenn O'Donnell said: "There's really no stopping this juggernaut. They will continue to dominate."