NVIDIA just reported its strongest quarter ever, with $82 billion in revenue and a surprising new product that could reshape how AI systems work. The chip maker's latest earnings reveal that demand for its Blackwell architecture GPUs has accelerated beyond expectations, but executives are already pointing to the next frontier: agentic AI, a new category of AI systems that can autonomously complete complex tasks without constant human direction. To capture this emerging market, NVIDIA introduced Vera, its first custom CPU built specifically for agentic AI workloads, positioning the company to tap into what management estimates is a $200 billion total addressable market for CPUs.

What Is Agentic AI and Why Does It Matter?

Agentic AI represents a fundamental shift in how artificial intelligence operates. Unlike today's chatbots that respond to individual prompts, agentic AI systems can set goals, break down complex problems into steps, and execute tasks autonomously across multiple applications. Think of it as the difference between asking a calculator for a single answer versus hiring an employee who can plan a project, delegate subtasks, and report back with results. NVIDIA management stated that "demand has gone parabolic. The reason is simple. Agentic AI has arrived. AI can now do productive and valuable work. Tokens are now profitable, so model makers are in a race to produce more". This shift is already driving infrastructure spending at an unprecedented scale.

How Does Vera Position NVIDIA for the Next Wave of AI?

Vera and its paired platform, VeraRubin, represent NVIDIA's bet that agentic AI requires different hardware than the GPUs that power today's large language models. While Blackwell GPUs excel at processing massive amounts of data in parallel, agentic AI workloads demand different performance characteristics: lower latency, higher energy efficiency, and the ability to coordinate multiple tasks simultaneously. NVIDIA expects Vera to generate approximately $20 billion in standalone CPU revenue this year alone, with production shipments of VeraRubin beginning in the third quarter and ramping through subsequent quarters. The company is positioning Vera as a catalyst product that will enable new system architectures optimized for autonomous AI agents.

The introduction of Vera also signals NVIDIA's confidence that the CPU market, traditionally dominated by Intel and AMD, represents an untapped opportunity for AI-specialized processors. By designing a CPU from the ground up for agentic workflows rather than general-purpose computing, NVIDIA is attempting to replicate the success it achieved with GPUs in the AI era.

What Are the Key Drivers Behind NVIDIA's Record Quarter?

While Vera captures the strategic vision, NVIDIA's immediate growth is being fueled by extraordinary demand for Blackwell GPUs across multiple customer segments. The company's data center business generated $75 billion in revenue, up 92 percent year over year, with Blackwell accounting for the majority of shipments. This growth is being driven by several distinct customer categories:

• Hyperscale Cloud Providers: Companies like Amazon, Google, and Microsoft operating massive data centers generated $38 billion in revenue, representing approximately 50 percent of total data center revenue and growing 12 percent sequentially.
• AI Cloud, Industrial, and Enterprise Customers: A newly defined segment capturing non-hyperscale deployments, including sovereign AI infrastructure and regional cloud providers, generated $37 billion in revenue and grew 31 percent sequentially, with AI cloud revenue more than tripling year over year.
• Networking Infrastructure: NVIDIA's InfiniBand networking products, which connect GPUs in data centers, more than quadrupled in revenue year over year, bolstered by next-generation XDR deployments that enable faster communication between chips.

The breadth of this growth across different customer types suggests that AI infrastructure spending is not concentrated in a handful of mega-cap tech companies but is spreading across enterprises, governments, and regional cloud providers worldwide. NVIDIA noted that sovereign AI revenue, which represents infrastructure deployed in countries building their own AI capabilities, grew more than 80 percent year over year and now spans nearly 40 countries.

How Is NVIDIA Reshaping Its Business to Reflect Market Changes?

NVIDIA made a significant organizational change by re-segmenting its data center business into Hyperscale and ACIE (AI Cloud, Industrial, and Enterprise) subsegments. CEO Jen-Hsun Huang explained that this restructuring "aims to capture the distinct needs and go-to-market dynamics of Hyperscale, ACIE, and Edge platforms to give investors clearer visibility into growth drivers". This shift reflects a maturing AI infrastructure market where different customer types have fundamentally different requirements, pricing dynamics, and deployment timelines.

The company also reported that the number of partner data centers exceeding 10 megawatts of power capacity nearly doubled in one year, now exceeding 80 sites globally. This metric signals the massive infrastructure buildout underway to support AI workloads, as companies invest billions in facilities capable of housing thousands of GPUs and managing the enormous power consumption required to run them.

What Does NVIDIA's Guidance Tell Us About Future Growth?

NVIDIA guided for $91 billion in revenue for the next quarter, representing continued sequential growth of approximately 11 percent. The company expects gross margins to remain stable at approximately 75 percent, indicating that despite the massive scale of production, Blackwell remains highly profitable. Operating expenses are expected to grow in the upper 40 percent range for the full year, driven by higher research and development spending and increased usage of AI productivity tools internally.

One notable caveat: NVIDIA explicitly excluded China data center compute revenue from its outlook, reflecting caution amid continued export licensing uncertainty for H200 GPU shipments to mainland customers. This represents a significant headwind, as China has historically been a major market for NVIDIA's data center products. The company's decision to provide conservative guidance on China suggests that geopolitical tensions around AI chip exports remain a material risk to future growth.

Why Are Investors Focused on Vera When Blackwell Is Already Booming?

While Blackwell GPUs are driving current revenue growth, Vera represents NVIDIA's strategic positioning for the next phase of AI development. The company has emphasized that agentic AI, not just larger language models, is the next frontier. Management noted that "LPX and similar SRAM-based accelerators are expected to remain niche, reaffirming Blackwell and Vera platforms as central to the company's growth strategy". This statement signals confidence that the company's core product roadmap is aligned with where the market is heading.

Management

The introduction of Vera also addresses a potential vulnerability in NVIDIA's business model. If agentic AI workloads truly require different hardware characteristics than current GPU-centric architectures, then competitors like Intel, AMD, or custom silicon designers could capture significant market share. By introducing Vera now, NVIDIA is attempting to own the agentic AI hardware category before competitors can establish a foothold.

NVIDIA's record quarter and forward guidance demonstrate that the company has successfully transitioned from a graphics chip maker to the central infrastructure provider for the AI era. The introduction of Vera signals that NVIDIA is not resting on its GPU dominance but is actively positioning itself to lead the next wave of AI hardware innovation. For investors and industry observers, the real question is whether agentic AI will deliver on its promise of autonomous, productive work, or whether it remains a speculative category. NVIDIA's $20 billion revenue bet on Vera suggests the company believes the former is far more likely.