Nvidia has shifted its strategic focus beyond GPU sales to architecting and deploying a global AI ecosystem that spans data centers, energy grids and software layers. The company is investing in modular software stacks for inference and training, while collaborating with utility providers to integrate AI-optimized servers directly into power distribution networks. This holistic approach aims to drive recurring infrastructure revenue streams and create high switching costs, positioning Nvidia as a critical partner for enterprises looking to scale AI workloads at grid-scale performance and efficiency.

Nvidia is in advanced discussions to acquire AI21 Labs for a deal valued between $2 billion and $3 billion, up from AI21’s last known $1.4 billion valuation in 2023. The startup’s 200-strong team of AI researchers and engineers is seen as a key asset to accelerate Nvidia’s development of large-scale language models. AI21 paused work on its consumer writing assistant earlier this year, and joining Nvidia would provide the resources to integrate those capabilities into the broader AI stack. Completion of the deal could be announced in Q1 2026, pending regulatory approvals.

Nvidia reports visibility into nearly $500 billion in combined Blackwell and Rubin platform contracts from the start of 2025 through the end of 2026, of which $150 billion has already been fulfilled. Production of the next-generation Vera Rubin platform is on schedule for ramp in H2 2026, combining a new CPU architecture with advanced GPUs to support cloud, enterprise, robotics and physical AI workloads. This extended backlog underpins analyst models projecting double-digit revenue growth and margin expansion through 2027.