Prabowo Widodo
Jensen Huang, the visionary founder and CEO of Nvidia, a figure often hailed as one of the most effective corporate evangelists of our time, has once again captivated the technology world, asserting that Nvidia has unlocked a "brand new $200 billion Total Addressable Market (TAM)" with its recently introduced Vera CPU. This bold proclamation, delivered during the company’s latest earnings call, follows yet another quarter of record-breaking financial performance, reinforcing Huang’s reputation for delivering on ambitious promises. His relentless optimism, which some might compare to Salesforce’s Marc Benioff, is consistently backed by Nvidia’s impressive execution and market dominance in the rapidly expanding artificial intelligence sector.
Nvidia’s Unstoppable Momentum and the AI Revolution
Nvidia’s financial trajectory has been nothing short of meteoric, driven primarily by the insatiable demand for its Graphics Processing Units (GPUs) in the burgeoning AI landscape. The company reported a staggering $81.6 billion in revenue for the recent quarter, exceeding analyst expectations, and further projected an astounding $91 billion for the upcoming quarter. These figures underscore Nvidia’s central role in the global AI infrastructure build-out, solidifying its position as a technological titan. The company’s market capitalization has soared, reflecting investor confidence in its continued innovation and market leadership.
However, even as Nvidia achieves unprecedented success, Wall Street analysts and industry observers consistently harbor anxieties about potential threats to its seemingly unassailable perch. The primary concern has historically revolved around the CPU market, a domain traditionally dominated by established players like Intel and AMD. While Nvidia has made previous forays into CPUs, such as with its Tegra line and the more recent Grace CPU, these ventures have not been its core business, which has remained firmly rooted in high-performance GPUs. The landscape of AI hardware, however, is evolving rapidly, prompting companies across the spectrum to explore new architectures and specialized chips.
The Strategic Introduction of Vera: Targeting Agentic AI
It is within this dynamic environment that Nvidia strategically introduced Vera, its new CPU product, in March. Huang positioned Vera as a potentially transformative product, specifically engineered for the emerging paradigm of "agentic AI." Speaking on Wednesday’s earnings call, following the revelation of Nvidia’s blockbuster financial results, Huang articulated a compelling vision for Vera, highlighting its promising early sales figures and its potential to become a major growth driver for the company.
The term "agentic AI" refers to autonomous AI entities or software agents designed to perform specific tasks, interact with environments, and make decisions independently. Unlike traditional AI models that might primarily focus on data processing or pattern recognition, agentic AI systems are envisioned as proactive entities capable of complex reasoning, planning, and execution. Huang believes that these agents will proliferate, becoming as ubiquitous as human users on PCs today, and will require a specialized processing architecture.
Unlocking a New TAM: The Agentic AI Imperative
Huang’s declaration of a "brand new $200 billion TAM" is rooted in the unique processing demands of agentic AI. He elaborated on the distinction between the "thinking" part of an AI model, which predominantly relies on GPUs for parallel processing of vast datasets, and the "doing" or operational aspect, which he asserts will primarily run on CPUs. These agents, he predicts, will effectively run their own form of CPU-driven "PCs," necessitating a different kind of central processing unit.
Vera is designed precisely for this purpose. Huang explained that Vera is "the world’s first CPU, purpose-built for agentic AI," engineered to process "tokens as fast as possible." This design philosophy stands in contrast to classic cloud architecture CPUs, which are typically optimized with multiple "cores" for running numerous instances of applications concurrently and efficiently handling general-purpose computing tasks. For agentic AI, the speed and efficiency of sequential token processing are paramount, as agents interpret instructions, make decisions, and execute actions in a step-by-step, interactive manner. This specialization, Huang contends, provides Nvidia with a distinct competitive advantage in this nascent but rapidly expanding market.
"Vera opens a brand new $200 billion TAM for Nvidia, a market we have never addressed before, and every major hyperscaler and system maker is partnering with us to deploy it," Huang stated on the call, emphasizing the broad industry adoption already underway. "The world is rebuilding computing for agentic AI and robotic physical AI. Nvidia sits at the center of these transitions." This statement highlights not only the commercial opportunity but also Nvidia’s strategic positioning at the forefront of the next wave of AI evolution.
The Competitive Gauntlet: Hyperscalers and In-House Chips
Despite Huang’s confident assertions and Nvidia’s formidable track record, the competitive landscape for AI chips is intensifying. Major cloud providers and numerous startups are heavily investing in developing their own custom AI silicon, aiming to reduce reliance on external vendors, optimize performance for their specific workloads, and potentially lower costs.
A notable example of this trend emerged recently when Amazon Web Services (AWS) announced a significant contract with Meta for millions of Amazon’s homegrown AI CPUs. AWS CEO Andy Jassy has been vocal about his belief that AWS can develop AI chips, encompassing both GPUs and CPUs, that are at least as competitive, and potentially superior, to those offered by market leaders like Nvidia and Intel. This aggressive stance from a hyperscale giant like AWS underscores the existential threat that in-house chip development poses to traditional semiconductor suppliers. Similarly, Google has its Tensor Processing Units (TPUs), and Microsoft is investing heavily in its own AI accelerators.
For Nvidia, venturing into the CPU market, even with a specialized product like Vera, means directly confronting these powerful entities on their home turf. The historical dominance of Intel and AMD in the CPU space provides a formidable baseline, while the hyperscalers’ internal development capabilities represent a significant and growing challenge.
Nvidia’s Credibility and Early Market Validation
However, Huang’s claims are not merely speculative. He provided tangible evidence of Vera’s early market acceptance, revealing that Nvidia has already sold $20 billion worth of standalone Vera CPUs this year. This substantial figure, achieved relatively early in the product’s lifecycle, lends considerable weight to his predictions and suggests that Nvidia’s partners and customers are buying into the vision for agentic AI.
Huang’s long-term vision extends far beyond current applications. He anticipates a future where billions of AI agents operate autonomously, requiring vast amounts of specialized CPU power. "The world has a billion users, human users. My sense is that the world is going to have billions of agents, not today. I mean, we’re going to grow into it, but we’ll have billions of agents, and those billions of agents will all use tools. And those tools are going to be like PCs, just like us humans using PCs today," he elaborated. "We’re going to need a lot more CPUs." This analogy to the personal computer revolution underscores the potential scale of the agentic AI market and Nvidia’s ambition to be its foundational hardware provider.
Technical Superiority and Strategic Bundling
The technical differentiation of Vera is central to Nvidia’s strategy. By designing a CPU specifically to accelerate token processing, Nvidia aims to achieve performance efficiencies that general-purpose CPUs cannot match for agentic workloads. This specialization, combined with Nvidia’s established ecosystem of software tools, libraries, and developer support (CUDA platform), creates a powerful value proposition. Furthermore, Vera is sold both as a standalone unit and bundled with Nvidia’s flagship Rubin GPU, offering customers integrated solutions that leverage Nvidia’s full-stack expertise. This bundling strategy not only provides a comprehensive solution for AI workloads but also strengthens Nvidia’s overall ecosystem lock-in.
Broader Implications and Future Outlook
The implications of Nvidia’s push into agentic AI CPUs are profound. If Huang’s predictions materialize, it would fundamentally reshape the AI hardware market, creating a new segment where Nvidia could establish early dominance. This move would also diversify Nvidia’s revenue streams, making it less solely reliant on GPU sales for training large AI models and positioning it for the inference and operational phases of AI deployment.
For competitors, Vera represents a significant challenge. Intel and AMD will need to respond with their own specialized CPU designs or risk being marginalized in this high-growth segment. Hyperscalers developing their own chips will face an even more formidable competitor in Nvidia, which is now actively innovating across the entire AI hardware stack. The race for AI supremacy is not just about raw processing power; it’s increasingly about specialized architectures tailored for specific AI paradigms.
Ultimately, Nvidia’s foray into agentic AI CPUs with Vera is a testament to its aggressive pursuit of growth and its ability to anticipate and capitalize on emerging technological shifts. While skepticism is a natural reaction to such bold claims, Huang’s consistent track record of turning audacious visions into market realities has earned him, and Nvidia, a significant degree of trust from investors and the industry alike. The $200 billion question is not if agentic AI will transform computing, but how quickly, and whether Nvidia’s Vera CPU will indeed become its indispensable engine.
