$2B Investment: NVIDIA Backs Marvell Technology

According to recent industry reports, NVIDIA has invested approximately $2 billion in Marvell Technology to enable customers to more easily develop customized artificial intelligence (AI) chips using Marvell’s capabilities in combination with NVIDIA’s networking infrastructure and central processing technologies.

Through this strategic investment, NVIDIA aims to reinforce its central role in addressing the rapidly growing computational demands of AI applications. At the same time, an increasing number of companies are turning to custom processors as alternatives to NVIDIA’s higher-cost general-purpose solutions, signaling a shift in the competitive landscape of AI silicon.

According to Jacob Bourne, the collaboration gives NVIDIA access to Marvell’s semi-custom chip expertise and advanced optical interconnect technologies. These capabilities are critical for scaling data center–level AI systems, where bandwidth and energy efficiency remain key bottlenecks. Bourne noted that the move also broadens NVIDIA’s ecosystem to include a wider range of specialized silicon, helping the company maintain its position as a primary gateway for increasingly diverse AI workloads.

As Marvell strengthens its foothold in AI data center technologies, the partnership is expected to benefit from strong market momentum. Bourne added that investors may view the deal as a friction-reduction strategy, allowing third-party AI chips to operate more seamlessly within NVIDIA-dominated data center environments. This approach could enable NVIDIA to preserve its leadership while simultaneously expanding the scope and adoption of AI semiconductor solutions.

Both companies plan to jointly develop advanced networking solutions for AI, with a focus on optical interconnects and silicon photonics technologies. These innovations are designed to deliver high-speed, energy-efficient data transmission—key requirements for next-generation AI infrastructure.

Marvell will provide customized chips and networking solutions compatible with NVIDIA’s NVLink Fusion architecture, while NVIDIA will contribute complementary technologies, including CPUs, network interface cards (NICs), and interconnect platforms.

Major technology companies such as Alphabet Inc. and Meta Platforms are expected to collectively invest at least $630 billion in AI infrastructure this year. This surge in capital expenditure is projected to drive demand for chips used in servers and networking equipment, benefiting companies like Marvell.

Marvell also stated that it expects its fiscal 2028 revenue to grow by nearly 40%, reaching close to $15 billion, supported by sustained demand from AI-driven data center deployments.