Rochan Sankar's $900 Million Move to Nvidia: A Turning Point in GPU Infrastructure

Nvidia just made a major power move in the AI infrastructure space by recruiting Rochan Sankar, Enfabrica’s chief executive, along with his engineering team and securing full access to the company’s cutting-edge technology stack. According to CNBC’s reporting, the transaction, finalized last week, was structured as a combination of cash and equity compensation. This acquisition marks a significant escalation in Nvidia’s strategy to consolidate AI talent and critical infrastructure capabilities under one roof.

What Makes This Deal Different

What Nvidia is essentially acquiring isn’t just personnel—it’s the technology that allows massive GPU clusters to operate as a unified system. Enfabrica’s infrastructure layer, which can orchestrate 100,000 graphics processors to function cohesively, represents exactly the kind of vertical integration Nvidia needs. Founded in 2019, Enfabrica built its reputation by solving a fundamental problem in AI systems: how to make GPU clusters communicate seamlessly at scale. Microsoft’s Wisconsin data center project, which deploys 72 GPUs working in tandem within individual racks, exemplifies the infrastructure problem Enfabrica was designed to solve.

The timing is strategic. Nvidia had already committed capital during Enfabrica’s Series B funding round in 2023, when Atreides Management led a $125 million investment. At that juncture, the startup’s valuation had increased fivefold from the previous round. This latest move represents Nvidia converting its investor relationship into operational control.

The Broader Acquisition-Hiring Playbook in AI

Rochan Sankar’s transition to Nvidia echoes a pattern that’s become the dominant hiring strategy among tech giants navigating AI expansion. Meta deployed $14.3 billion to bring Alexandr Wang, founder of Scale AI, into its organization while acquiring a 49% ownership stake. Google executed a comparable maneuver, absorbing Varun Mohan from Windsurf in a $2.4 billion transaction that included technology licensing arrangements. The Character.AI team integration at Google, Microsoft’s acquisition of Inflection, and Amazon’s pickup of Adept all follow the same template: identify emerging AI infrastructure or talent, package it as an acquisition, and avoid triggering antitrust scrutiny that would accompany traditional company purchases.

Nvidia’s Measured Approach to Acquisitions

Despite dominating the AI chip landscape, Nvidia has been remarkably selective with major acquisitions. The company’s only billion-dollar-plus deal prior to recent activity occurred in 2019, when it paid $6.9 billion for Mellanox, an Israeli chip design firm. That networking infrastructure remains central to Nvidia’s Blackwell architecture. An attempted $40 billion acquisition of Arm Corporation stalled when regulators blocked the transaction in 2022, effectively teaching Nvidia that full company purchases face regulatory headwinds.

The $700 million acquisition of Run:ai, a Tel Aviv-based infrastructure optimization platform, fits the pattern of Nvidia building operational control over the software layer that sits atop its hardware. Beyond direct acquisitions, Nvidia has deployed capital strategically: a $5 billion position in Intel announced this week, coupled with commitments to jointly develop AI processing systems, and a $700 million investment in Nscale, a British data center technology firm.

The Scale of Nvidia’s Market Evolution

Nvidia’s valuation trajectory illustrates why these infrastructure moves matter. The company crossed $4.28 trillion in market capitalization recently, representing a fourfold increase since 2023. This valuation expansion reflects market recognition that Nvidia’s competitive advantage extends beyond chip design into the entire ecosystem required to deploy AI infrastructure at scale. Each acquisition of infrastructure companies, teams, and technology represents another layer of moat around its core GPU business.

The Rochan Sankar recruitment encapsulates Nvidia’s evolution from a chip manufacturer into a comprehensive AI systems provider—a transition that requires controlling not just processors but the entire technical stack required to make those processors perform at their theoretical maximum across massively distributed deployments.