Xsight Labs' $2.8B Bet on Open Networking to Power the AI Arms Race

TEL AVIV, Israel – July 30, 2026

In a resounding vote of confidence for the future of open data center infrastructure, fabless semiconductor firm Xsight Labs has closed a funding round of over $300 million, catapulting its valuation to $2.8 billion. The investment, led by Fidelity Management & Research Company, provides the Israeli-American startup with a formidable war chest to challenge industry incumbents in the high-stakes race to build the networking backbone for artificial intelligence and next-generation cloud services.

Founded in 2017, the company has rapidly gained traction with its E1 Data Processing Unit (DPU) and X2 Ethernet switch, which are already being adopted by global network operators and evaluated by the world's largest hyperscale cloud providers. This new capital is poised to accelerate a direct challenge to the closed, proprietary systems that have long dominated the networking landscape, offering a glimpse into a more flexible, efficient, and powerful digital future.

Beyond the Walled Garden: The Industry's Pivot to Openness

The massive valuation reflects a fundamental shift occurring deep within the world's data centers. For decades, network operators were often locked into single-vendor ecosystems, where hardware and software were tightly integrated. These "closed, legacy architectures" could stifle innovation, slow down deployment cycles, and inflate costs. Today, the explosive growth of AI and cloud computing demands a level of scale, speed, and flexibility that these traditional models struggle to provide.

This demand has fueled a powerful movement toward open networking, championed by initiatives like the Open Compute Project (OCP) and Microsoft's Software for Open Networking in the Cloud (SONiC). SONiC decouples the network operating system from the underlying hardware, allowing operators to run the same software across equipment from multiple vendors. The industry's adoption has been swift. In March 2024, Gartner predicted that 40% of organizations with large data center networks would run SONiC in production by 2025. Market analysts at 650 Group have projected the SONiC-enabled data center switching market could exceed $9 billion by 2028, expanding far beyond its hyperscaler origins. This shift is particularly critical for AI, where open, disaggregated systems enable the massive, low-latency, and lossless fabrics needed for efficient communication between thousands of GPUs.

A Challenger in a Field of Giants

Xsight Labs is stepping onto a field dominated by titans. NVIDIA, with its integrated Spectrum-X platform, saw its data center revenues surge 43% year-over-year in the first quarter of 2026 on the back of AI demand. Likewise, established players like Cisco, Broadcom, and Arista Networks continue to command significant market share. Yet, Xsight Labs has carved out a distinct and compelling niche. The company claims to be the only semiconductor firm currently shipping both high-end DPUs and Ethernet switches that are fully programmable and built on a foundation of open standards.

"This valuation is a testament to the team's relentless execution and reflects the market need for a high-performance alternative to closed, legacy architectures," said Yossi Meyouhas, CEO of Xsight Labs, in the company's announcement. The startup's E1 DPU is reportedly the first 800G DPU to ship and, crucially, the first to complete SONiC-DASH validation at that speed. DASH (Disaggregated APIs for SONiC Host-interface) is a project that standardizes how applications offload networking tasks to a DPU. This validation is more than a technical badge; it's a guarantee to hyperscalers that the hardware will integrate seamlessly into their open-source environments, drastically reducing deployment risk and engineering overhead.

Fueling the AI Infrastructure Arms Race

The investment from Fidelity and other major players is more than a bet on a single company; it's a strategic move in the global AI infrastructure arms race. As enterprises and cloud providers rush to build out massive AI clusters, the demand for underlying network components has skyrocketed. According to LightCounting, sales of the Ethernet optical transceivers that connect these systems are expected to double in 2024 alone. Meanwhile, IDC reported that 800G switches, the speed offered by Xsight's products, already accounted for nearly 36% of data center segment revenues in the first quarter of 2026, cementing them as the new standard for AI builds.

In this environment, performance-per-watt has become a critical metric. The immense power consumption and heat generated by AI clusters pose significant operational and environmental challenges. Xsight Labs' technology directly addresses this, with its E1 DPU delivering a reported 4x improvement in performance per watt over previous designs. By offering a combination of raw speed, programmability, and extreme power efficiency, the company is positioning its silicon as a foundational building block for the sustainable scaling of AI.

From Capital to Execution: The Challenge of Scale

With $300 million in fresh capital, Xsight Labs now faces the formidable challenge of execution. The semiconductor industry is notorious for its supply chain volatility, and securing manufacturing capacity at advanced foundries is a fiercely competitive endeavor. Furthermore, meeting the high-volume, mission-critical demands of Tier-1 hyperscalers requires a global logistics and support operation of immense scale and precision. Industry-wide bottlenecks are already causing long wait times for key components, with some customers waiting until 2025 for certain optical transceiver deliveries.

Xsight Labs has earmarked its new funding to tackle these hurdles head-on. The capital will be used to accelerate its multi-generational product roadmap, including support for the emerging Ultra Ethernet Consortium (UEC) standards. It will also fuel a significant expansion of its engineering and customer support teams across the US, Israel, Europe, and Asia. Most importantly, the funds will be deployed to scale manufacturing and supply chain operations, a direct investment in ensuring the company can fulfill the high-volume orders its technology is now attracting. This strategic preparation is critical as the startup moves from a promising innovator to a core supplier in the engine room of the global digital economy.