The Intelligence Toll: Why Every Fortune 500 Company Could Pay Nvidia by 2035
Nvidia's stock, currently valued at 40 times forward earnings, could see significant growth if artificial general intelligence (AGI) arrives by 2030. Analysts predict revenue could reach $1 trillion by 2035, leading to a market cap of $9 trillion. The company's competitive edge lies in its CUDA software, which has a strong lock-in effect. However, risks include potential competition from AMD and geopolitical tensions. The investment outlook hinges on the timely arrival of AGI, with bullish targets suggesting a stock price increase to between $369 and $615, while bearish scenarios could see it drop to $150-$200.
At 40 times forward earnings, Nvidia (NVDA 0.60%) looks expensive through a traditional semiconductor lens. But that framework collapses if artificial general intelligence (AGI) arrives by 2030, as OpenAI, Anthropic, and other labs privately expect. Nvidia won't just supply artificial intelligence (AI) infrastructure. It could collect a toll on every intelligent operation on the planet.
Think of it as the intelligence toll: a per-cycle fee on every autonomous decision, every simulated experiment, every AI agent running across every industry. When computing becomes as essential as electricity, Nvidia becomes the billing system for intelligence itself.
Image source: Getty Images.
The path to doubling
The math is simpler than skeptics think. Nvidia's revenue hit $130.5 billion in fiscal 2025, more than doubling from the prior year. Wall Street already expects $254 billion by fiscal 2027. But that's just the beginning. The real explosion could come if AGI transforms every industry.
A compound growth rate of 19% from 2027 to 2035 gets you to $1 trillion in revenue. At 45% net margin and an earnings multiple of 20, reasonable for critical infrastructure, that's a $9 trillion market cap. With 24.39 billion shares outstanding, that translates to $369 per share, a double from today.
Bull case? If Nvidia captures 50% of a $5 trillion AGI computing market, the stock could reach $615. Not the 20x return some imagine, but doubling your money as a company grows from $4 trillion to $9 trillion is hardly settling.
Fantasy? Microsoft committed $100 billion to AI data centers. Saudi Arabia, the United Arab Emirates, and Japan pledged $90 billion for sovereign computing. OpenAI alone spends $7 billion annually on Nvidia hardware. The smart money isn't betting on chatbots. It's prepaying for infrastructure in the AGI economy.
The trillion-dollar question
The entire investment case hangs on one question: Will AGI arrive by 2030? Answering it correctly is the difference between owning a stock that merely grows with the semiconductor industry and one that doubles as it becomes the world's first $9 trillion company. This chasm exists because current models only see the present: GPU sales for AI training. They fail to price in the explosion that happens when AGI becomes the engine of every industry.
Imagine pharmaceutical companies simulating every possible drug interaction, enterprises deploying millions of autonomous AI workers, and billions of users adopting personal AI companions. The first signals are already here. Salesforce reports more than 8,000 customers already using its AI agents, but the infrastructure required to power this future will dwarf today's entire market.
The CUDA moat
Nvidia's competitive moat wasn't dug overnight. It was built over 15 years by an army of 2 million developers, making its Compute Unified Device Architecture (CUDA) software the native language of AI. The cost to leave this ecosystem is a multibillion-dollar tax paid in retraining models (a $100 million expense for a GPT-4 class model), porting years of code, and accepting major performance penalties.
The proof is that even the world's richest tech giants, after spending billions building their custom chips, still buy Nvidia's GPUs. Alphabet pays the toll. Tesla pays the toll. This powerful lock-in is why every serious contender in the race to AGI -- OpenAI, Anthropic, xAI -- has aligned with Nvidia. They aren't just choosing a vendor; they're choosing the only battle-tested stack.
But even the strongest fortress can have a vulnerability. The moat is deepest around the specialized work of training AI. For the high-volume world of AI inference, rival armies from Advanced Micro Devices and cloud giants are gathering at the gates, building alternative routes that bypass the CUDA toll.
The risk of thinking small
Yes, a Taiwan invasion could crater the stock 70%. Custom application-specific integrated circuits (ASICs) might compress margins. AGI could arrive late. But here's the real calculation:
| Factor | Bull Case ($369 to $615 Target) | Bear Case ($150 to $200 Target) |
|---|---|---|
| Market role | Planetary-scale intelligence utility | Excellent but standard semiconductor leader |
| Margin | 70% or higher gross margin sustained | Margin compresses to 30% to 50% because of competition |
| AGI timeline | Arrives by 2030, driving exponential demand | Slips to 2040 or later, or never arrives in this form |
| CUDA moat | Remains the dominant standard | Open source and ASICs create viable alternatives |
| Market share | Captures 30% to 50% of the multitrillion-dollar AGI market | Loses share to AMD, hyperscalers, custom chips |
Nvidia isn't just riding the AI cycle. It's laying the rails for AGI itself. When every company needs AI like they need electricity, when every decision runs through neural networks, when intelligence becomes a utility, that's when early investors realize they didn't buy a semiconductor stock. They bought the cognitive infrastructure toll of the 21st century.
