Jensen Huang stood on stage at Nvidia’s GTC conference in March and laid out a vision so ambitious it bordered on audacious: a company already worth $2.6 trillion positioning itself to capture even more of the artificial intelligence infrastructure buildout. The stock has already returned more than 2,000% in the past five years. And yet, according to at least one prominent forecast, Nvidia could soar another 120%, pushing the chipmaker past the $5 trillion market capitalization mark.
That’s not a typo.
The case, outlined in detail by The Motley Fool, rests on a confluence of factors: Nvidia’s dominance in data center GPUs, the accelerating capital expenditure plans of hyperscale cloud providers, and the company’s expanding reach into new product categories that didn’t exist three years ago. It also rests on a belief that the AI infrastructure cycle is nowhere close to peaking — a belief that, if wrong, would make the bull thesis collapse spectacularly.
But first, the numbers.
Nvidia reported fiscal year 2025 revenue of $130.5 billion, a staggering 114% increase over the prior year. Data center revenue alone hit $115.2 billion. The company’s gross margins have hovered near 73%, a figure that would be considered extraordinary in almost any hardware business. Net income roughly tripled. Free cash flow surged. By nearly every financial metric that matters, Nvidia is executing at a level that makes its mega-cap peers look sluggish.
The bull case for a doubling from current levels hinges on what comes next: the Blackwell architecture and its successors. Nvidia’s Blackwell GPUs, which began shipping in volume in late 2024, represent a generational leap in AI training and inference performance. Management has indicated that Blackwell demand is “incredible” — Huang’s word — and that the company is working to ramp production as fast as supply chains allow. The next-generation Rubin architecture is already in development, with Vera Rubin expected to follow. Nvidia is now on an annual release cadence for its GPU platforms, a pace more reminiscent of smartphone chipmakers than traditional semiconductor companies.
The Motley Fool’s analysis points to a scenario where Nvidia’s revenue could approach or exceed $280 billion within the next few years, driven primarily by continued data center GPU sales but supplemented by newer business lines. At a forward price-to-earnings multiple of roughly 25 to 30 times — reasonable for a company growing this fast — that math gets you to a $5 trillion market cap without requiring heroic assumptions about margin expansion.
Here’s where it gets interesting. The $1 trillion question, as The Motley Fool frames it, isn’t really about Nvidia’s chip performance. It’s about whether the customers buying these chips — Microsoft, Google, Amazon, Meta, Oracle, and a growing list of sovereign AI initiatives — will sustain their current spending trajectory. The hyperscalers have collectively signaled more than $300 billion in capital expenditure for 2025 alone, with a significant portion directed toward AI infrastructure. Microsoft has committed roughly $80 billion. Meta has indicated plans north of $60 billion. Amazon and Google are in similar territory.
These are enormous sums. Unprecedented, actually, for a single technology cycle.
The skeptics — and there are thoughtful ones — argue that this level of spending cannot be sustained without a corresponding surge in AI-driven revenue for these companies. So far, the returns on AI investment have been promising but uneven. Microsoft’s Azure AI services are growing rapidly. Google’s AI-enhanced search and cloud offerings are gaining traction. But the revenue generated by AI applications hasn’t yet matched the capital being deployed to build the underlying infrastructure. If the hyperscalers pull back, Nvidia’s growth rate decelerates fast.
That concern isn’t theoretical. In previous technology cycles — fiber optics in the late 1990s, cloud computing buildouts in the mid-2010s — infrastructure spending ran ahead of demand before correcting. The AI cycle could follow a similar pattern. Or it might not. The difference this time, bulls argue, is that AI workloads are genuinely transformative across industries: drug discovery, autonomous vehicles, software development, financial modeling, content generation. The breadth of application is wider than any prior computing platform shift.
Nvidia’s competitive position makes the bull case more defensible than it would be for a lesser franchise. The company controls an estimated 80% to 90% of the data center GPU market for AI workloads. Its CUDA software platform, built over nearly two decades, creates switching costs that competitors have struggled to overcome. AMD’s MI300 series has gained some traction, particularly with cost-conscious buyers. Intel’s Gaudi accelerators have found niche applications. And custom silicon from Google (TPUs), Amazon (Trainium and Inferentia), and other hyperscalers is a real and growing competitive factor.
But none of these alternatives have dented Nvidia’s market share in a meaningful way. Not yet.
The custom chip threat deserves more scrutiny than it typically receives. When your largest customers are also developing their own competing products, the long-term dynamics get complicated. Amazon’s Trainium2 chips are being deployed at scale within AWS. Google has been using TPUs internally for years and is now offering them to external cloud customers. Microsoft is developing its own AI accelerator, Maia. These efforts won’t replace Nvidia GPUs entirely — the workloads are too diverse and Nvidia’s software advantages too deep — but they could cap Nvidia’s pricing power and market share over time.
Nvidia’s response has been to expand beyond chips. The company’s networking business, bolstered by its 2020 acquisition of Mellanox, is now a multi-billion-dollar revenue stream. InfiniBand and Ethernet networking products are critical for connecting thousands of GPUs in large-scale AI clusters. Nvidia’s NVLink interconnect technology allows GPUs to communicate with each other at speeds that conventional networking can’t match. As AI models grow larger and training clusters scale to tens of thousands of GPUs, this networking layer becomes increasingly valuable.
Then there’s software. Nvidia has been building out a portfolio of enterprise AI software, including its NIM (Nvidia Inference Microservices) platform and the Omniverse simulation environment. These products are still relatively early in their revenue contribution, but they represent Nvidia’s ambition to capture value across the entire AI stack — not just at the silicon layer. Software carries higher margins and creates recurring revenue streams, two characteristics that investors reward handsomely.
The automotive and robotics segments add another dimension. Nvidia’s DRIVE platform for autonomous vehicles and its Isaac platform for robotics are positioned to benefit from the next wave of AI deployment beyond the data center. Huang has spoken extensively about “physical AI” — the application of AI models to robots, self-driving cars, and industrial automation. These markets are earlier stage but potentially enormous. If humanoid robots become a commercial reality in the next decade, as companies like Tesla, Figure, and Agility Robotics are betting, Nvidia’s computing platforms could be at the center of that market.
So what could prevent Nvidia from reaching $5 trillion?
Geopolitics, for one. The U.S. government’s export controls on advanced AI chips to China have already cost Nvidia billions in potential revenue. China represented a significant portion of Nvidia’s data center sales before the restrictions tightened. While Nvidia has developed lower-performance chips designed to comply with export rules, the regulatory environment remains fluid and could become more restrictive. A broader trade conflict or additional sanctions could further limit Nvidia’s addressable market.
Valuation compression is another risk. Nvidia currently trades at roughly 25 to 30 times forward earnings, depending on which analyst estimates you use. That’s not cheap in absolute terms, but it’s reasonable relative to the company’s growth rate. If growth slows — whether because of a demand pullback, increased competition, or macroeconomic headwinds — the multiple could contract significantly. A stock trading at 20 times earnings looks very different from one trading at 30 times, even if the underlying business is still growing.
Execution risk shouldn’t be dismissed either. Nvidia is attempting to ramp multiple new product lines simultaneously while managing a global supply chain that depends on Taiwan Semiconductor Manufacturing Company for its most advanced chips. Any disruption at TSMC — whether from natural disaster, geopolitical tension over Taiwan, or simple manufacturing yield issues — would ripple through Nvidia’s business immediately.
And there’s the broader macro picture. Interest rates, while off their 2023-2024 peaks, remain elevated by the standards of the previous decade. Higher rates increase the discount rate applied to future earnings, which mathematically reduces the present value of high-growth stocks. A recession — or even a meaningful economic slowdown — could cause enterprise customers to delay AI investments, creating a demand air pocket that Nvidia would feel acutely.
Still, the weight of evidence tilts bullish for now. Wall Street consensus estimates project Nvidia’s revenue growing by more than 50% in fiscal year 2026, with continued strong growth in fiscal 2027 as Blackwell ramps fully and new product cycles begin. Analyst price targets from major banks range widely but cluster in a zone that implies meaningful upside from current levels. Morgan Stanley, Goldman Sachs, and Bank of America have all maintained bullish stances on the stock, citing the durability of AI infrastructure spending and Nvidia’s competitive moat.
The comparison to historical precedents is instructive but imperfect. Nvidia’s rise has drawn parallels to Cisco during the internet buildout of the late 1990s — a company that supplied the critical infrastructure for a transformative technology wave. Cisco peaked at roughly $555 billion in market cap in March 2000 and then lost 80% of its value over the following two years. The analogy is seductive but flawed in key respects: Nvidia’s profitability is far superior to Cisco’s at a comparable stage, its competitive advantages are more durable, and the AI workload cycle appears to have a longer runway than the initial internet infrastructure buildout.
That said, no analogy is perfect. And past performance guarantees nothing.
What makes the $5 trillion scenario plausible — not certain, but plausible — is the sheer scale of the opportunity Nvidia is addressing. The total addressable market for AI infrastructure, including chips, networking, software, and services, is estimated by various industry analysts at $500 billion to $1 trillion annually by the end of the decade. If Nvidia captures even 30% to 40% of that market, the revenue base would support a valuation well north of where the stock trades today.
The company has also demonstrated an ability to expand its addressable market in ways that weren’t obvious even two years ago. Sovereign AI — the push by national governments to build domestic AI computing capacity — has emerged as a meaningful new demand driver. Countries including Saudi Arabia, the UAE, India, Japan, France, and Singapore have announced plans to invest billions in AI infrastructure, much of it built on Nvidia hardware. This is incremental demand that barely existed in 2023.
Inference is another growth vector that’s still in its early innings. While Nvidia’s initial AI revenue surge was driven primarily by training large language models, the inference market — running those trained models at scale to serve billions of users — is expected to grow even larger over time. Nvidia’s GPUs are well-suited for inference workloads, and the company’s software optimizations for inference continue to improve. As AI applications proliferate across consumer and enterprise settings, inference compute demand should scale accordingly.
Jensen Huang has described the current moment as the beginning of a new industrial revolution — one driven by AI rather than steam or electricity. That’s the kind of statement a CEO makes when he wants investors to think in decades rather than quarters. Whether you buy the analogy or not, the financial results so far are hard to argue with. Nvidia isn’t projecting future dominance. It’s demonstrating current dominance while arguing the market is about to get much bigger.
For investors weighing whether Nvidia can double from here, the question ultimately isn’t about the company’s technology or competitive position — both are formidable. It’s about the sustainability of the AI spending cycle and whether the returns on that spending will justify continued investment. If the answer is yes, Nvidia at $5 trillion looks not just possible but logical. If the answer is no, the correction will be swift and painful, as it always is when capital-intensive buildouts outrun demand.
The market, as usual, will be the final arbiter. But Nvidia has earned the right to be taken seriously when it makes big claims. Few companies in the history of the semiconductor industry have executed this well, this consistently, at this scale. Whether that execution can push the stock to $5 trillion is the most consequential bet in technology investing right now.
WebProNews is an iEntry Publication