LAN Infrastructure Management: More Assisted Than Automated

In enterprise LAN infrastructures, AI is driving the offering forward, not the demand.

Just over a year ago, we summarized the market situation as depicted by the Magic Quadrant for this space in that way. It was a kind of shorthand. Gartner did not establish an explicit link between AI and demand. It argued mainly that the pullback in demand was due to longer product life cycles, a consequence of economic pressures.

The US firm, however, did connect AI to the development of the offering. It stated, in that context, that it was propelled by new procurement approaches amid a shift toward autonomous networks. Among these approaches, NaaS. Still only modestly adopted (less than 2% of companies), it was nevertheless gradually taking shape, with consumption models based for example on space occupied or on per-user increments.

Gartner relegates NaaS to the background

This year, NaaS sits noticeably more in the background in the Magic Quadrant’s synthesis for LAN infrastructures. Gartner hardly mentions it, save for the very end of the list of “optional” criteria (evaluated but not essential to be included in the ranking).

From year to year, the list of “mandatory” criteria remains fairly tight. In broad terms:

• Unified management
• Lifecycle automation
• Access security
• Isolation and microsegmentation
• Streaming telemetry
• ITSM integrations
• Public API
• Switches with PoE and MACsec encryption
• Tri-band Wi‑Fi access points (with 6 GHz)

Last year’s optional criteria, ZTNA remained optional. UEBA was as well. As well as, among others:

• Automated quarantine
• Digital twins
• AI agents to process telemetry and day-to-day operations
• Natural language assistant for configuration and troubleshooting
• Detection and remediation of configuration drift
• Indoor localization
• Data residency controls
• Management of third-party equipment
• Flexible licensing models (including NaaS)
• AI-enabled radio optimization
• Generative UI for operations

Fortinet is no longer the “leader”; Cisco regains it

The Magic Quadrant is structured around two axes: “Execution” and “Vision.” The former reflects the ability to meet demand (quality of products/services, customer experience, business viability, etc.). The latter reflects strategy (geography, industry focus, product, go-to-market, marketing…).

The situation on the “Execution” axis:

On the “Vision” axis:

Most offerings include AI-assisted diagnostics, correlations and recommendations. Beyond these building blocks increasingly treated as a baseline, closed-loop automation is currently developing for specific use cases; it is not yet transversal. Adoption of agent-based solutions remains limited, focused on validating changes, incident analysis and detecting configuration drift.

Segmentation and access control based on identities are widely available, but integrating them without increasing administrative overhead remains challenging. Differences between vendors show up on this point, and more broadly on the quality of implementation. All the more so in a context where AI speeds up the delivery of features… and potentially accelerates fragmentation.

HPE and Huawei, again “leaders”

Juniper, a “leader” in 2025, no longer appears in the Magic Quadrant this year. The consequence of its acquisition by HPE. The latter, already a “leader” last year with the Aruba offering, reinforces its position.

Another “leader” in 2025 that remains so: Huawei. Cisco, by contrast, returns to this circle after being downgraded to “challengers” among other reasons because its product strategy was not well aligned with needs.

A fourth vendor is a “leader”: Arista Networks. Gartner had classified it as a visionary for several years.

Aside from Juniper, only one vendor has left the Magic Quadrant: TP-Link (formerly a niche player). Its offering nonetheless earns it an “honorable mention.”

Arista Networks keeps the datacenter label

Gartner highlights for Arista Networks:

• Operational consistency around the CloudVision platform, which includes NDR
• Use of AI for diagnostics, simulation (digital twins) and change validation
• Financial stability and a growing customer base

It notes, however:

• A lack of branding in this segment, buyers often associating Arista Networks with the datacenter universe
• A product range largely geared toward large enterprises, impacting pricing
• Adoption mainly in North America, which may limit consistency across other regions

Between Catalyst and Meraki, Cisco hasn’t clarified everything

At Cisco, Gartner praises:

• Increasingly coherent operation between the Catalyst and Meraki portfolios
• The shared foundation IOS-XE + Deep Network Model
• The scale of the business, the customer base and the partner ecosystem

It notes, however:

• Rising purchasing complexity, especially when switching between Meraki and Catalyst
• A lack of clarity on the positioning of Meraki and Catalyst within the offerings during evaluations
• Uneven AI delivery, some elements not yet covering the full portfolio

Mist or Central? At HPE, a portfolio choice

In 2025, Gartner had praised HPE for:

• The integration of Aruba Networking Central and OpsRamp, easing migrations
• AI features, especially around telemetry collection and processing
• The fabric-based approach that facilitates zero-trust deployments

It had, at the same time, pointed out:

• Functional disparities between cloud and on-site offerings
• The multiplicity of NaaS options
• Uncertainties related to the Juniper acquisition

Positive points this year:

• Alignment with demand for AIOps
• Maturity of Aruba Central and Mist, and the engine they share
• Financial investments, broad customer base and the channel

Negative points:

• Need to choose between Mist and Central, with operational implications
• In this context, new customers tend to be more cautious about validating the roadmap
• Increased effort to evaluate solutions

Huawei remains a geopolitical issue

In 2025, Huawei stood out for:

• AI capabilities (digital twins, provisioning UX optimization)
• The fabric approach that facilitates zero-trust deployments
• The roadmap aligned with needs (cloud/on-prem parity, integration of Wi‑Fi 8 concepts into Wi‑Fi 7 access points…)

Gartner nonetheless highlighted:

• Commercial presence tied to geopolitics
• Limited effectiveness of a marketing strategy oriented toward technology
• Reluctance to offer opex models

Positive points this year:

• Closed-loop automation
• Unified CloudCampus architecture, with AIOps and zero trust
• Rapid feature delivery cadence, notably in Wi‑Fi detection, AI-assisted support and cybersecurity automation

Points to watch:

• Regulatory and geopolitical restrictions again
• Also again, a techno-centric positioning that may resonate less in business-centric evaluations
• Regional model variability, likely to create uneven purchasing and support experiences

Fortinet, focused on its firewall

Downgraded from “leader” to “visionary,” Fortinet is currently implementing AI in a diagnostic assistance role rather than a full automation logic. It has a control plan centered on its firewall, which may limit autonomous evolution of LAN operations. More broadly, its offering is positioned as a component of its cybersecurity portfolio, which can reduce visibility during evaluations.