A Citi report notes that NVIDIA has announced a new $150 billion share buyback authorization, bringing its total buyback capacity to $235 billion by FY28, signaling confidence in AI-driven free cash flow growth. Meanwhile, the company is expanding from GPUs into AI security, software, and full-stack infrastructure, with revenue per GW expected to rise from $18 billion in the Hopper era to $40 billion in the Rubin era.
$NVIDIA (NVDA.US)$It is evolving from a GPU‑chip company into a full‑stack platform provider spanning computing power, networking, security, and AI infrastructure, as the AI wave simultaneously expands both the scale of demand and the per‑project value.
Citi recently released a report, maintaining its "Buy" rating on NVIDIA with a price target of $315. The report noted that NVIDIA has announced an additional $150 billion share buyback authorization, bringing its total buyback capacity through FY28 to $235 billion, reflecting the company's confidence in the continued growth of AI-driven free cash flow.
Meanwhile, NVIDIA is expanding into AI safety, compliance, governance, and comprehensive AI factory infrastructure, while the rapid proliferation of open-weight models continues to drive up token consumption and underlying computing power demand.
More notably, NVIDIA is increasing the value of each AI infrastructure deployment. The company expects the revenue opportunity per gigawatt to rise from $18 billion in the Hopper era to $25 billion in the Blackwell era, and further to $40 billion in the Rubin era.
$150 Billion Share Repurchase: Cash-Flow Confidence Behind the Massive Authorization
Following NVIDIA's announcement of a new $150 billion share buyback authorization, its total buyback capacity through FY28 has reached $235 billion. With the company's market capitalization standing at $5.52 trillion, this scale of buybacks has become a key component of its capital allocation strategy.
Citi believes that the further expansion of the buyback program reflects NVIDIA's confidence in the future growth of its free cash flow. As demand for AI infrastructure continues to expand, the company expects its profitability and cash-generating capacity to remain robust, enabling it to sustain substantial shareholder returns while continuing to increase its investments in the AI sector.
In other words, NVIDIA has not had to choose between investing for growth and delivering returns to shareholders. On the one hand, the company continues to invest in AI infrastructure, software, and emerging applications; on the other hand, it has also maintained ample room for share repurchases.
Extending from GPUs to AI Security: Business Boundaries Continue to Expand
NVIDIA has launched the Open Agent Safety Platform, marking its entry into the safety, compliance, and governance aspects of AI agents. The platform includes the OpenShell open-source runtime and the Sentry hardware monitoring system, which runs on the BlueField-4 DPU. It enables policy-based control over AI agents and detects and blocks unauthorized activities.
As AI agents integrate into enterprises' critical workflows, companies' demand for security and control capabilities is rising in tandem. Accordingly, NVIDIA is further extending its reach from underlying computing to the security layers required for running AI applications, thereby expanding its addressable market.
Meanwhile, the open-model ecosystem is expanding rapidly. Currently, open models account for about 75% of token generation, up from roughly 40% a year ago; overall token consumption has grown 25-fold year over year. NVIDIA believes that increasing model openness is driving the widespread adoption of AI applications, while rising token demand is further boosting underlying computing power requirements.
The company continues to invest in areas such as language AI, physics‑based AI, robotics, and autonomous driving, and plans to acquire Hugging Face to strengthen its ties with the developer ecosystem.
The revenue opportunity per gigawatt has increased from $18 billion to $40 billion.
NVIDIA expects that the revenue opportunity per GW has increased from $18 billion in the Hopper era to $25 billion in the Blackwell era, and will reach $40 billion in the Rubin era.
This growth stems not only from improved GPU performance but also from the inclusion of CPUs, LPUs, networking equipment, and AI‑factory infrastructure in its end‑to‑end solutions. As a result, NVIDIA can cover more stages in AI infrastructure deployment, thereby increasing the revenue scale of individual projects.
Order data also shows that demand for AI infrastructure continues to expand. NVIDIA disclosed that a cutting-edge AI laboratory has already directly contracted for 2.6 GW of NVIDIA AI infrastructure, with the relevant equipment scheduled for delivery before 2028; in addition, the combined indirect contract value of several cloud service providers (CSPs) and emerging cloud vendors has exceeded US$180 billion.
This means that NVIDIA's growth stems not only from the expanding scale of AI infrastructure deployments, but also from the company's ability to cover more products and stages within a single project. As Hopper evolves into Blackwell and Rubin, the revenue opportunity per gigawatt continues to rise.
Editor/Deng