Nvidia Eyes $1 Trillion Revenue With New Chips

Nvidia’s CEO Jensen Huang has set an audacious target: as much as $1 trillion in revenue over time, driven by a new wave of data‑centre products. Huang highlighted Rubin and Blackwell accelerators alongside a newly revealed Vera CPU and updated server rack designs as the core elements of that strategy.

According to Huang, Rubin and Blackwell are intended to scale AI training and inference workloads across hyperscale and enterprise environments. The Rubin architecture focuses on efficiency and throughput for mixed AI tasks, while Blackwell aims at top-tier performance for large‑scale models. Together, they form a chip roadmap Nvidia expects will capture a large slice of cloud and on‑premise compute demand.

Beyond GPUs, Nvidia is pushing into general‑purpose CPU territory with Vera — a server CPU effort positioned to work closely with the company’s accelerators. Huang portrayed Vera as part of a broader systems play: integrated servers and racks that reduce latency, raise utilization and simplify deployment for customers running complex AI pipelines.

Investors and customers watching the announcement are weighing a few factors: how quickly data‑centre customers will adopt a combined Nvidia compute stack, and how competitors and chip ecosystems respond. Nvidia’s advantage lies in its software stack and established presence in AI workloads, but broader CPU and system adoption will require ecosystem support and validated performance claims in real deployments.

For readers tracking enterprise AI infrastructure, this marks Nvidia’s continued shift from selling discrete GPUs to offering full hardware‑software solutions. If Rubin, Blackwell and Vera deliver on promises, Nvidia could solidify a more entrenched position across cloud providers and enterprises alike — though execution and market dynamics will determine the ultimate payoff.