By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
Panelists are cautious about Nvidia's stock valuation and potential risks in the face of impressive growth and a $1 trillion opportunity. They highlight the risk of hyperscalers shifting to custom chips for inference, which could commoditize Nvidia's offerings and squeeze margins.
Risk: Hyperscalers shifting inference to internal silicon, potentially commoditizing Nvidia's offerings and squeezing margins.
Opportunity: The $1 trillion opportunity in Blackwell/Rubin systems, expanding the total addressable market.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Nvidia's AI demand is expanding beyond training into reasoning and agentic AI workloads.
Blackwell and Rubin systems could extend Nvidia's growth cycle far longer than many investors expect.
AI monetization is accelerating at an unprecedented pace.
Artificial intelligence (AI) has been the most prominent investment theme on Wall Street over the past few years. Nvidia (NASDAQ: NVDA) has benefited dramatically from this trend, with the stock up over 640% in the past three years.
But with its next earnings report on May 20, investors are concerned about whether the stock has already run too far or if there is still more upside left.
Will AI create the world's first trillionaire? Our team just released a report on the one little-known company, called an "Indispensable Monopoly" providing the critical technology Nvidia and Intel both need. Continue »
Here are some reasons why I am considering buying the stock before the numbers come out.
Nvidia has guided for revenue of about $78 billion, plus or minus 2%, for the first quarter of fiscal 2027 (ending April 26, 2026). That implies roughly 73% to 80% year-over-year growth, which is an exceptionally strong growth rate for a company of Nvidia's size.
The company's recent performance already reflects strong momentum. Nvidia's revenue soared 73% year over year to $68.1 billion in the fourth quarter of fiscal 2026. The company's data center business generated revenue of $62.3 billion, up 75% year over year.
While Nvidia has already positioned itself as an AI infrastructure company, the changing nature of AI demand suggests that the growth opportunity may be even larger than it appears today.
Management has highlighted that AI is evolving from content creation to reasoning and now toward agentic AI, where systems can independently perform tasks. Since these systems need to continuously think, read information, reason, and generate outputs, they require significantly more inference computing capacity. Subsequently, power-constrained data centers operate like "token factories," continuously generating AI output, or tokens. Instead of focusing solely on chip costs, customers are increasingly evaluating how many tokens their systems can generate per unit of power.
Nvidia's latest product cycle is focused on addressing these evolving AI workloads. The company's Blackwell systems are already seeing strong demand. Previously, management had highlighted $500 billion worth of high-confidence demand and purchase orders tied to Blackwell and next-generation Rubin systems through 2026. However, recently, CEO Jensen Huang said that he expects to see at least $1 trillion of opportunity tied to these systems through 2027. Since that forecast excludes additional opportunities from stand-alone CPUs, storage, and recently licensed Groq inferencing (running AI models in a production environment) technology, the actual total addressable market could be even larger.
The Rubin system is expected to deliver performance improvements far beyond traditional chip upgrades, especially for more advanced AI tasks such as reasoning and agentic AI. With Nvidia delivering large performance gains by combining chips, networking, and software into complete systems, customers see improvement in the economics of their AI deployments. That is helping support Nvidia's strong revenue growth and profit margins.
Nvidia is also becoming more aggressive about securing AI infrastructure capacity directly. The company plans to invest up to $2.1 billion in data center operator Iren as part of a partnership to deploy up to 5 gigawatts of AI infrastructure.
Nvidia is also investing in the AI infrastructure supply chain. The company is helping fund new factories for glassmaker Corning through a multibillion-dollar prepayment. Corning's glass is used in fiber-optic cables required for networking infrastructure inside AI data centers.
While the top five hyperscalers account for nearly 60% of Nvidia's business, the remaining 40% is from enterprises, sovereign AI projects, regional clouds, industrial applications, robotics, big systems, supercomputing systems, small servers, and edge computing. The diversified customer base makes Nvidia resilient to spending slowdowns from any single industry or group of customers.
AI monetization also appears to be improving faster than expected. Management highlighted that some AI-native companies are reportedly adding nearly $1 billion to $2 billion in revenue every week as AI adoption increases. That helps address one of the biggest concerns around AI spending: whether customers can eventually generate meaningful returns from these investments.
Management has also highlighted that inference is becoming critical because it directly drives customer revenue. As AI systems handle more reasoning, coding, search, and agentic workloads, companies need significantly more computing capacity to generate tokens and serve users efficiently.
These trends underline the durable, broader, and more commercial nature of AI demand.
Nvidia is exposed to risks such as export restrictions in China, competitive pressures from chip designers and hyperscalers developing proprietary chips, reduced AI spending, and a high valuation. Despite these challenges, the company's broader growth story remains intact.
Going into the May 20 earnings report, investor expectations are undoubtedly high. But Nvidia's underlying demand drivers still appear strong, broad-based, and increasingly commercial.
If management's long-term demand commentary proves correct, Nvidia's share price can soar even higher in the next few years.
Before you buy stock in Nvidia, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Nvidia wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $471,827! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,319,291!
Now, it’s worth noting Stock Advisor’s total average return is 986% — a market-crushing outperformance compared to 207% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of May 10, 2026. *
Manali Pradhan, CFA has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Corning and Nvidia. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"Nvidia’s transition to an inference-heavy 'token factory' model is structurally sound, but the current valuation ignores the inevitable deceleration risks inherent in scaling at this magnitude."
Nvidia’s pivot from training to inference and agentic workflows is the correct structural narrative, but the market is pricing in perfection. While the $1 trillion opportunity in Blackwell/Rubin systems is compelling, investors are ignoring the law of large numbers. Maintaining 70%+ growth against a $78 billion revenue base requires an unprecedented expansion of the total addressable market that assumes zero meaningful cooling in hyperscaler CapEx. The reliance on 'token factory' metrics is a clever way to mask potential hardware saturation. I am cautious because the stock’s current valuation leaves zero margin for error if supply chain bottlenecks or sovereign AI funding delays manifest in the Q1 print.
If inference demand truly scales linearly with agentic AI adoption, Nvidia’s current valuation is actually a discount, as the market is significantly underestimating the long-term recurring revenue potential of their software-defined hardware ecosystem.
"Nvidia's growth remains robust but decelerating, with omitted China risks, rising ASIC competition, and rich valuation offering scant margin for error pre-earnings."
Nvidia's Q4 FY2026 revenue hit $68.1B (up 73% YoY), with data center at $62.3B (up 75%), and Q1 FY2027 guidance of $78B (±2%) implies 73-80% growth—impressive for its scale but decelerating from prior triple-digit surges. The article hypes $1T 'opportunity' for Blackwell/Rubin as locked demand, but it's vague TAM excluding CPUs/storage, while omitting China export bans (potentially 15-20% revenue hit) and hyperscaler shift to custom chips (Google TPUs, Amazon Trainium). At undisclosed but likely 35x+ forward P/E amid capex bloat from Iren/Corning investments, perfection is priced in for May 20 earnings.
AI's pivot to power-hungry inference/agentic workloads uniquely favors Nvidia's full-stack systems (chips+networking+software), with $1T pipeline and diversified 40% non-hyperscaler demand enabling sustained 50%+ growth and margin expansion.
"Management's $1T Blackwell/Rubin forecast is aspirational guidance, not contracted revenue—the real test is whether hyperscaler capex remains rational or enters bubble territory by late 2026."
The article conflates *demand signals* with *realized demand*. Yes, Nvidia guides 73-80% YoY growth and management cites $1T opportunity through 2027—but this is forward-looking commentary, not booked revenue. The inference-driven 'token factory' thesis is real, yet the article glosses over a critical risk: if hyperscalers (60% of revenue) hit ROI walls or pull back capex, the diversified 40% won't absorb the shock fast enough. Also, the $2.1B Iren investment and Corning prepayments are *capital deployment*, not demand validation—they're Nvidia securing its own supply chain, which suggests internal hedging against uncertainty.
Nvidia's valuation already prices in this $1T thesis. If Q1 FY2027 misses guidance by even 3-5 points, or if management walks back the Rubin TAM, the stock re-rates sharply lower despite the underlying business remaining strong.
"Nvidia's rally hinges on an outsized, durable AI monetization ramp from agentic workloads; if that ramp proves weaker or shorter-lived than the article implies, the upside risks are substantial."
Article frames Nvidia as riding an AI demand surge into reasoning and agentic workloads, with Blackwell/Rubin expanding TAM and a bold Q1 FY2027 guide of about $78B, implying 73-80% YoY growth. If true, the stock could re-rate. Yet the bullish case rests on highly optimistic monetization and a multi-year data-center capex cycle that may decelerate. The piece glosses over whether real demand converts to durable profits, and it underplays risks like China export controls, regulatory review, competition from closed AI stacks, and potential CAPEX peaks. A data point miss or weaker guidance could trigger meaningful multiple compression even with solid growth.
Even if the demand thesis holds, Nvidia's stock is pricing in perfection; near-term upside could be limited if earnings momentum slows or AI spending normalizes earlier than expected. The thesis hinges on an unsustainably large multiplier on an evolving AI spend cycle.
"The rise of internal custom silicon for inference will commoditize Nvidia's hardware and erode its premium pricing power."
Grok, your focus on custom silicon like Google’s TPU is the real sleeper risk. While Nvidia dominates the training cluster, inference is becoming a commodity race where hyperscalers will prioritize cost-per-token over Nvidia’s full-stack premium. If AWS and Google successfully shift inference to internal silicon, Nvidia’s 'token factory' narrative collapses into a pure hardware margin squeeze. We are ignoring the shift from 'must-have' proprietary chips to 'good-enough' internal alternatives that prioritize balance-sheet efficiency over peak performance.
"Nvidia's CUDA software moat and sovereign AI diversification blunt hyperscaler custom silicon and China risks."
Gemini, your inference commoditization fear overlooks Nvidia's CUDA moat: hyperscalers can't swap to TPUs without massive retraining costs and ecosystem lock-in breakage, as agentic AI demands seamless low-latency inference. Grok's 15-20% China hit is overstated—recent bans target high-end chips, but legacy sales persist. Sovereign AI (enterprise/sovereign now 40%+ of DC revenue) accelerates independently of hyperscaler capex wobbles Claude notes.
"Inference commoditization risk is real if hyperscalers achieve 80%+ performance parity at materially lower cost—CUDA switching costs matter less when the workload is price-elastic."
Grok's CUDA lock-in defense is overstated. Hyperscalers don't need perfect parity—they need 85% of Nvidia's performance at 60% cost. Inference workloads are latency-tolerant compared to training; TPU/Trainium adoption curves suggest they're already crossing that threshold. The real question: does Nvidia's 40% non-hyperscaler revenue grow fast enough to offset a 10-15% hyperscaler inference shift? Nobody's modeled that scenario.
"Rapid internal hyperscaler silicon adoption could erode Nvidia's margins and trigger a re-rating unless pricing power and non-hyperscaler growth hold up."
Gemini, your inference commoditization concern is valid, but the bigger risk is pace. If hyperscalers accelerate internal silicon adoption faster than anticipated, Nvidia could face margin compression even with steady unit growth, as pricing power wanes and service costs rise. CUDA/eco lock-in is not a permanent moat in a race to lower $/token. A sustained 40% non-hyperscaler mix may not fully offset a sharp hyperscaler capex pullback or accelerated chip deprecation.
Panelists are cautious about Nvidia's stock valuation and potential risks in the face of impressive growth and a $1 trillion opportunity. They highlight the risk of hyperscalers shifting to custom chips for inference, which could commoditize Nvidia's offerings and squeeze margins.
The $1 trillion opportunity in Blackwell/Rubin systems, expanding the total addressable market.
Hyperscalers shifting inference to internal silicon, potentially commoditizing Nvidia's offerings and squeezing margins.