Nvidia has quietly closed its largest talent acquisition to date, investing over $900 million to bring Enfabrica’s networking expertise and leadership into its AI infrastructure fold. The deal targets Enfabrica CEO Rochan Sankar, key engineering talent, and a license to the startup’s breakthrough high-scale networking technology. This strategic “acquihire” positions Nvidia to deliver next-generation, rack-scale AI systems and reinforces its dominance in a fierce war for top AI talent.
Enfabrica, founded in 2019 by Sankar and Shrijeet Mukherjee, built its reputation solving one of AI’s most pressing challenges: keeping hundreds of GPUs communicating at full speed. Its flagship SuperNIC and network-on-chip solutions can link more than 100,000 GPUs without performance bottlenecks, dramatically reducing idle time and increasing throughput. Nvidia’s recent Blackwell rack systems—housing up to 72 GPUs per enclosure—stand to benefit immensely from this technology.
Under the agreement, Nvidia pays a mix of cash and equity to enroll Enfabrica’s core team while securing perpetual licensing rights to its networking innovations. Sankar joins Nvidia as Senior Vice President of Data Center Networking, bringing 25+ years of industry experience, 11 patents in networking and memory, and a proven track record from Broadcom’s Ethernet switching division. His leadership will accelerate Nvidia’s shift from standalone GPUs toward integrated AI supercomputers.
Industry analysts view this deal as part of a rising trend of acquihires—high-value, talent-focused transactions that avoid regulatory hurdles tied to full mergers. Meta kickstarted the wave with its $14.3 billion investment in Scale AI, while Google licensed tech from Windsurf for $2.4 billion. With the FTC scrutinizing Big Tech consolidations, skill-driven deals offer a nimble pathway to secure scarce expertise.
Enfabrica’s prior funding history underscores its rise: $260 million raised to date, including a $115 million Series B in late 2024 valuing the startup at $600 million. Backers included Advanced Micro Devices, highlighting the technology’s cross-industry potential. This financial backing allowed Enfabrica to develop its SuperNIC chip, which cuts total cost of ownership for large-scale AI clusters by optimizing packet routing directly on the hardware.
Nvidia’s timing is critical. Amid China’s ban on domestic firms using Nvidia’s AI GPUs, the company must innovate to sustain growth. By integrating Enfabrica’s networking IP, Nvidia can offer turnkey AI systems that rival bespoke hyperscale deployments. This tighter hardware-software synergy could unlock new enterprise and cloud partnerships, cementing Nvidia’s leadership as demand for generative AI accelerates.
For website publishers and AI practitioners, the Nvidia–Enfabrica deal signals the expanding role of networking in AI performance. The move also underscores the premium placed on senior talent with semiconductor IP portfolios. As competition intensifies, expect more high-value acquihires shaping the future of AI infrastructure.
Read also:
