By: Jeff Gorman

At a moment when artificial intelligence is reshaping global markets, geopolitical calculations, and the architecture of modern computing itself, Nvidia’s Israeli operations have emerged as one of the most consequential—yet often understated—engines driving that transformation. What began as a strategic acquisition more than half a decade ago has matured into a sprawling research-and-development network that now stands second only to the United States in importance within Nvidia’s global ecosystem. From networking silicon and optical interconnects to AI infrastructure software and next-generation computing platforms, Israel has become indispensable to Nvidia’s vision of the future.

Today, Nvidia’s Israeli network employs more than 5,000 people, with thousands more expected to join as new campuses rise and acquisitions are integrated. It is not merely a regional outpost or auxiliary innovation lab. Rather, it is a foundational pillar of the company’s AI strategy, shaping how massive data centers communicate, how workloads are orchestrated, and how future chips will power the next wave of computational demand.

The story of Nvidia in Israel cannot be told without returning to 2019, when the Silicon Valley giant acquired Mellanox Technologies for nearly $7 billion. At the time, the deal was widely viewed as a bold wager on high-performance networking—a domain that, while essential, operated largely behind the scenes of the computing world. Mellanox, headquartered in Yokneam, Israel, was already a leader in InfiniBand and Ethernet technologies that enabled ultra-fast, low-latency communication between servers.

That acquisition proved prescient. As artificial intelligence workloads exploded in scale, it became clear that raw compute power alone was insufficient. GPUs needed to communicate seamlessly across thousands—sometimes tens of thousands—of nodes. The bottleneck was no longer the processor; it was the network.

Mellanox’s technology, and the Israeli engineers behind it, became the connective tissue of Nvidia’s AI ambitions. What had once been a specialized networking company was now the backbone of the world’s most advanced AI supercomputers.

At the heart of Nvidia’s Israeli operation is a singular focus on networking and connectivity—fields that have become existential to AI at scale. Israeli teams are responsible for major components of Nvidia’s Spectrum-X Ethernet platform, a system designed to optimize Ethernet networks specifically for AI workloads. Unlike traditional enterprise networking, Spectrum-X is engineered to handle the intense east-west traffic patterns generated by large language models and deep neural networks.

Beyond Ethernet, Israeli engineers continue to advance InfiniBand, NVLink, and optical interconnect technologies, all of which are essential to linking GPUs within and across data centers. One of the most ambitious frontiers now under development involves connecting geographically remote data centers into what effectively functions as a single, massive AI machine. This requires breakthroughs in optical networking, latency management, and fault tolerance—areas in which Nvidia’s Israeli teams are deeply involved.

These innovations are not theoretical. They are already being deployed by hyperscalers, national research labs, and cloud providers racing to build AI infrastructure capable of supporting trillion-parameter models and real-time inference at global scale.

Perhaps the clearest indication of Israel’s centrality to Nvidia’s future lies in its role in the company’s next-generation computing platform, codenamed Rubin. Slated to succeed the current Grace-Blackwell architecture, Rubin represents Nvidia’s next leap forward in tightly integrated CPU-GPU systems optimized for AI and high-performance computing.

Israeli teams are contributing to multiple elements of this platform, including aspects of the Vera CPU and the advanced networking and switching technologies that will bind Rubin-based systems together. This level of involvement places Israel not merely at the periphery of Nvidia’s roadmap, but at its core—helping define the architectural principles that will govern AI computing for the remainder of the decade.

In practical terms, this means that decisions made in labs in Yokneam, Tel Aviv, and Be’er Sheva will influence how future data centers are designed, how energy efficiency is optimized, and how AI workloads are distributed across continents.

Nvidia’s Israeli footprint has expanded not only organically but through targeted acquisitions that strengthen its software and AI infrastructure stack. Among the most significant is Run:ai, an Israeli startup specializing in AI workload orchestration. As enterprises struggle to efficiently allocate GPU resources across multiple teams and applications, Run:ai’s software provides critical tools for maximizing utilization and reducing operational friction.

Another key acquisition is Deci, a company focused on AI model optimization. Deci’s technology allows models to run faster and more efficiently without sacrificing accuracy—an increasingly valuable capability as AI applications move from research labs into production environments with real-world constraints.

By integrating these companies into its Israeli network, Nvidia is reinforcing a holistic approach to AI infrastructure—one that spans hardware, networking, and software. Israel’s reputation for systems-level thinking and rapid innovation makes it a natural incubator for such cross-disciplinary development.

Nvidia’s confidence in its Israeli operations is perhaps most visible in its physical expansion. The company is investing billions of dollars in new campuses across northern Israel, while also deepening its presence in Be’er Sheva, a city that has positioned itself as a hub for cybersecurity and advanced engineering.

These campuses are designed not merely as office space, but as integrated innovation environments capable of supporting end-to-end development—from silicon design and firmware to networking protocols and AI software. The scale of the investment underscores Nvidia’s belief that Israel will remain central to its operations for decades to come.

At the same time, this expansion reflects a broader transformation of Israel’s tech ecosystem. Nvidia’s presence has become a magnet for talent, suppliers, and startups, reinforcing the country’s role as a global center for deep-tech innovation rather than consumer-facing applications alone.

Despite its success, Nvidia’s Israeli operation faces challenges, most notably a persistent shortage of skilled engineers. Competition for talent is fierce, not only among multinational corporations but also from startups and defense-related industries. Nvidia has responded by investing heavily in training, partnerships with academic institutions, and internal mobility programs that allow engineers to move between projects and disciplines.

Yet these pressures have not diminished Israel’s strategic value to Nvidia. On the contrary, the concentration of expertise in networking, systems engineering, and AI infrastructure has become increasingly difficult to replicate elsewhere. The Israeli network now plays a direct role in Nvidia’s revenue generation, as networking products and AI infrastructure solutions account for a growing share of the company’s business.

In the final analysis, Nvidia’s Israeli operations are not a satellite or a contingency—they are a core engine powering the AI age. From the chips that process data to the networks that move it at the speed of light, and the software that orchestrates it all, Israel’s imprint is embedded throughout Nvidia’s technology stack.

As AI systems grow larger, more distributed, and more integral to daily life, the importance of scalable, resilient infrastructure will only increase. Nvidia has bet that Israel’s unique blend of engineering rigor, entrepreneurial agility, and systems-level insight will remain essential to meeting that challenge.

That bet now appears less like a gamble and more like a strategic inevitability. In laboratories overlooking the Mediterranean and desert campuses in the Negev, the invisible architecture of the world’s AI future is being built—packet by packet, photon by photon, and line of code by line of code.