Arista Networks has officially launched its 7060XE7 Series, a comprehensive portfolio of 1.6T networking platforms purpose-built to underpin rack-scale AI infrastructure. This announcement marks a significant evolution from standalone high-performance switches to integrated rack-scale systems designed to handle the extreme density, power, and thermal efficiency required by modern AI workloads.

The 7060XE7 family comprises both fixed switch platforms and configurable rack-scale systems, all running Arista's Extensible Operating System (EOS). EOS features low-latency and intelligent packet buffering to manage the intense microbursts that are characteristic of AI communication and collective patterns. The systems are built on Broadcom’s Tomahawk 6 silicon, and Arista is also collaborating with AMD on next-generation compute silicon and NICs to enable scale-out AI fabrics.

Strategic Shift Toward Rack-Scale Systems

Strategically, the 7060XE7 Series signifies Arista’s transition from offering standalone switches to providing complete rack-scale systems. These platforms allow customers to build both scale-up and scale-out AI fabrics using air, liquid, and hybrid-cooled technology. The portfolio includes multiple configurations to suit different data center environments and deployment needs.

Specific Models and Availability

• 7060XE7-64PS and 7060XE7-64PRS (4U Rack Switches): Available in Q4 2026, these air-cooled systems support pluggable Integrated Heat Sink (IHS) and Riding Heat Sink (RHS) optics. IHS targets current air-cooled data centers, while RHS is aimed at future liquid-cooled AI fabrics and extreme port density.
• 7060XE7-64PRS-RV3-L: A specialized 2OU liquid-cooled platform for high-density clusters, featuring 224G SerDes. This system uses DC power from the ORv3 rack and contains no internal fans, integrating with liquid-cooled XPU servers to maximize power efficiency. Available in Q1 2027.
• 7060XE7-128PE: Also coming in Q1 2027, this device provides 128 800G ports in an air-cooled 4RU design, utilizing 100G SerDes for deployment flexibility and backward compatibility.

Software and Open Protocol Support

On the software side, EOS is the featured network operating system, but the family also supports open-source software such as SONiC (Software for Open Networking in the Cloud) and OpenSwitch. One of the portfolio’s key features is full support for the Open Compute Project’s Multipath Reliable Connection (MRC), an RDMA-based transport protocol that allows a single reliable connection to simultaneously use many network paths over Ethernet.

As described by Arista’s President and CTO Kenneth Duda and Distinguished Engineer Alan Judge, “MRC is an open protocol where endstation NICs stripe their traffic across multiple links and paths to the receiver, with out-of-order packets automatically handled. MRC responds to network congestion signals (ECN and packet trimming), shifting load to the best-performing paths, and avoiding links and paths that can’t actually reach the destination altogether.” MRC monitors each path, steering around congestion, avoiding paths with link errors, and avoiding failed links. This approach achieves very high fabric utilization with good load balancing while interoperating seamlessly with scale-across and WAN networks utilizing standard dynamic routing protocols.

The software also supports load balancing, congestion management, telemetry, and diagnostics—technologies that will be core to AI networking.

Market Context and Competition

The new Arista family joins a growing ecosystem of vendors looking to tap into the 1.6T Ethernet world, including Cisco, Nvidia, Celestica, and others. According to Sameh Boujelbene, Vice President of Data Center Switch and AI Networks Market Research at Dell’Oro Group, “Arista Network’s new 7060XE7 Series is a strong signal of where large-scale AI fabrics are heading: higher bandwidth, better power efficiency, and tighter integration between compute, optics, silicon, cooling, and network operating software.” She highlighted strong customer and ecosystem validation from Microsoft Azure, Oracle Cloud Infrastructure, Meta, AMD, and Broadcom.

Arista’s move into rack-scale systems reflects the broader industry trend toward tightly integrated infrastructure for AI. The ability to handle 1.6T per port enables higher aggregation speeds, reducing the number of tiers needed in AI clusters and lowering overall latency. The adoption of liquid cooling is also critical as power densities for AI servers continue to climb beyond what traditional air cooling can handle efficiently.

Background on Arista's Networking Legacy

Arista Networks, founded in 2004, has long been a leader in high-performance data center networking. The company’s EOS operating system is known for its modularity and programmability, making it a favorite among large-scale cloud providers and enterprises. With the 7060XE7 Series, Arista is doubling down on the AI infrastructure market, which demands unprecedented levels of bandwidth, low latency, and energy efficiency.

The Tomahawk 6 silicon from Broadcom is a key enabler, offering 51.2 Tbps switching capacity per chip. This allows Arista to deliver 64 ports of 800G or 128 ports of 400G, with future upgrades to 1.6T per port. The collaboration with AMD on compute and NICs ensures tight integration between the network and server components, optimizing end-to-end performance for AI workloads.

Future Outlook

As AI model sizes continue to grow, the demands on network infrastructure will only increase. Arista’s 7060XE7 Series positions the company to address these needs with a future-proof portfolio that scales from dense, air-cooled data centers to next-generation liquid-cooled clusters. The support for open standards like MRC and SONiC ensures interoperability and avoids vendor lock-in, a critical requirement for large-scale AI deployments.

The first models will start shipping in Q4 2026, with more advanced liquid-cooled and high-density versions arriving in early 2027. With strong backing from major cloud providers and a clear technology roadmap, Arista is well-positioned to capture a significant share of the growing AI networking market.

Source: Network World News