By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panel's net takeaway is that Nvidia's Vera Rubin platform could significantly impact AI hardware, but execution risks, customer responses, and potential commoditization pose challenges to the bullish case.
Risk: execution delays, customer inventory corrections, or saturation in hyperscaler capex cycles
Opportunity: supercharging the AI flywheel with cheaper AI boosting adoption and fattening hyperscaler margins
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 →
Key Points
Nvidia's new Vera Rubin data center chips for AI workloads offer more performance and better cost efficiency than its industry-leading Blackwell chips.
The Vera Rubin platform is likely to drive an acceleration in the company's revenue and earnings growth during its current fiscal year.
Nvidia stock could double, and potentially even triple, over the next two years.
- 10 stocks we like better than Nvidia ›
Nvidia (NASDAQ: NVDA) supplies the world's best graphics processing units (GPUs) for the data center, which are the primary chips used in artificial intelligence (AI) development. The company has an incredible amount of pricing power right now because demand continues to outstrip supply, which is driving a surge in its revenue and earnings.
Nvidia will start shipping its next generation of AI chips in the second half of this year. They are based on its new Vera Rubin architecture, which offers substantial improvements in performance and cost relative to its industry-leading Blackwell architecture.
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 »
Vera Rubin chips are expected to fuel an acceleration in Nvidia's revenue and earnings, which could translate into significant returns for the company's shareholders. Here's where I predict Nvidia stock will be in two years.
Vera Rubin could be Nvidia's most sought-after AI platform yet
Nvidia's GB300 GPU, which is based on its Blackwell Ultra architecture, is the most in-demand AI data center chip on the market right now. When configured in Nvidia's NVLink 72 data center rack, the GB300 offers up to 50 times more performance than the company's original AI data center chip, the H100, which was introduced in 2022. Greater performance can translate into more powerful AI models, and it can also lower costs by accelerating development timelines.
The Vera Rubin platform includes the Rubin GPU, the Vera CPU, the new NVLink 6 switches, and other networking hardware. Nvidia says it's so powerful that developers will be able to run the same AI training workloads with 75% fewer GPUs compared to Blackwell. Serving AI software to users will also be substantially cheaper, because Rubin reduces inference token costs by a staggering 90%.
Inference tokens are consumed when an AI model ingests an input from a user and generates an output in response. They can be words, images, or even videos, and they cost money to produce, which is why most AI companies charge their customers based on their usage rates.
Reducing inference token costs will achieve two things: First, it will make AI more affordable to use, thus boosting adoption, and second, it will improve profit margins for AI companies, giving them more money to fund additional infrastructure spending. This flywheel effect should, in theory, fuel more demand for Nvidia's chips over time.
Nvidia's growth is poised to accelerate
Nvidia delivered $215.9 billion in revenue during fiscal 2026 (ended Jan. 25), which was up 65% from the prior year. Its data center business alone brought in $193.7 billion, which was a 68% increase.
According to Wall Street's consensus estimate (provided by Yahoo! Finance), Nvidia could generate $367.7 billion in total revenue during fiscal 2027 (its current fiscal year), which would represent an accelerated growth rate of 70%. Most of that revenue is likely to come from the data center business yet again, given the incredible demand for AI chips.
At the bottom line, Wall Street analysts predict Nvidia's earnings will soar by 73% to $8.25 per share in fiscal 2027, accelerating from the company's fiscal 2026 earnings growth of 60%. Commercial quantities of Vera Rubin chips are expected to ship to customers in the second half of this calendar year, so they will play a central role in bringing analysts' forecasts to life.
Nvidia could be worth this much in two years
Based on Nvidia's fiscal 2026 earnings of $4.77 per share, its stock trades at a price-to-earnings (P/E) ratio of 37.2. That is a hefty discount to its 10-year average of 61.6, suggesting the stock might be undervalued right now.
The stock looks even cheaper on a forward basis. If we assume Nvidia will produce earnings of $8.25 per share during fiscal 2027 as Wall Street expects, then its forward P/E ratio is 21.8. Analysts believe the company will then deliver earnings of $10.80 per share in fiscal 2028, translating to a forward P/E ratio of 16.7.
Those forecasts suggest Nvidia stock will have to climb by 120% over the next two years just to maintain its current P/E ratio of 36.7, and by a whopping 269% if its stock were to trade in line with its 10-year average P/E of 61.6. Those potential gains imply a stock price of between $396 and $664 in two years from now, and an eye-popping market capitalization of between $9.6 trillion and $16.2 trillion.
Nvidia CEO Jensen Huang thinks AI infrastructure spending will reach $4 trillion per year by 2030, as data center operators race to meet demand for computing capacity from AI developers. Therefore, Nvidia stock could be poised for even more upside beyond the next two years.
Should you buy stock in Nvidia right now?
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 $508,877!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,115,328!*
Now, it’s worth noting Stock Advisor’s total average return is 936% — a market-crushing outperformance compared to 189% 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 March 19, 2026.
Anthony Di Pizio has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends 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
"Vera Rubin's technical superiority is real, but the article conflates product leadership with stock returns while ignoring competitive response and the risk that current valuation already prices in successful execution."
The article's math is internally consistent but rests on heroic assumptions. Vera Rubin delivering 75% fewer GPUs needed and 90% inference cost reduction is credible technologically—but the article assumes zero competitive response. AMD, Intel, and custom silicon (Google TPU, Meta's chips) are advancing too. The forward P/E compression from 21.8x (FY27) to 16.7x (FY28) assumes multiple contraction despite accelerating growth—historically Nvidia trades 40-60x forward during AI buildout phases. The $9.6-16.2T market cap by 2026 implies 2-3x upside, but that's priced in if Vera Rubin ships on schedule. The real risk: execution delays, customer inventory corrections, or saturation in hyperscaler capex cycles.
Nvidia's forward P/E of 21.8x is already below its 10-year average of 61.6x because the market is pricing in slower growth ahead or rising competition—not because it's cheap. If Vera Rubin disappoints or ships late, the stock re-rates lower, not higher.
"The article conflates technological superiority with financial valuation, ignoring that increased hardware efficiency often leads to lower unit volume demand for the manufacturer."
The article’s reliance on a 10-year average P/E of 61.6 to justify a $16 trillion market cap is intellectually dishonest. Nvidia is transitioning from a hyper-growth startup phase to a mature, albeit dominant, infrastructure giant. While the Rubin architecture’s efficiency gains are impressive, they create a 'deflationary' risk for Nvidia’s own top line: if you can do the same work with 75% fewer chips, the volume demand may not scale linearly with performance. At a forward P/E of 16.7, the market is clearly signaling skepticism regarding the sustainability of these margins. I am neutral; the growth is real, but the valuation math in this piece ignores the inevitable commoditization of AI hardware.
If the Rubin platform triggers an 'AI spring' by collapsing inference costs, the total addressable market could expand by an order of magnitude, rendering current P/E concerns irrelevant.
"Vera Rubin can materially boost Nvidia’s data-center economics, but doubling or tripling the stock in two years is far from guaranteed because it depends on execution, pricing power, and large TAM assumptions that are not certain."
The article’s bull case hinges on Vera Rubin delivering step-function improvements (the piece cites claims like 75% fewer GPUs for same training workloads and 90% lower inference token costs) and on Wall Street forecasts that push NVDA revenue toward $367.7B in fiscal 2027 with EPS of $8.25. Those inputs could certainly drive rapid re-rating, but the upside to the $396–$664 price targets requires flawless execution: timely high-volume shipments, maintained ASPs (average selling prices), and continued hyperscaler demand. Equally important are risks the article glosses over—customer verticalization, competition from custom accelerators, supply constraints, and a valuation that already embeds aggressive growth.
If Vera Rubin under-delivers on real-world total cost-of-ownership or hyperscalers accelerate their in-house chips, NVDA could miss the revenue/earnings ramps needed to justify current prices, sending the stock materially lower. High expectations also make NVDA vulnerable to even modest execution misses or macro slowdowns.
"Vera Rubin's efficiency edge positions NVDA for FY27 revenue acceleration to $368B and EPS re-rating from 22x to 30x+ forward multiples, driving 120-270% stock upside in two years."
Nvidia's Vera Rubin platform's claimed 75% fewer GPUs for training vs Blackwell and 90% inference cost reduction could supercharge the AI flywheel: cheaper AI boosts adoption, fattens hyperscaler margins, spurs more capex into NVDA data center chips (93% of FY26's $193.7B revenue). FY27 consensus $367.7B revenue (+70% YoY) and $8.25 EPS (+73%) at 21.8x forward P/E (vs 10-yr avg 61.6) screams undervaluation if Rubin ships on H2 schedule, targeting $396/share (120% upside) to maintain current multiple or $664 at historical norms for 9-16T market cap.
Hyperscalers like MSFT, AMZN, GOOG are ramping in-house ASICs (e.g., Azure Cobalt, Trainium) to cut NVDA dependency and pricing power, while AMD's MI300X chips gain traction; if AI capex ROI falters amid economic slowdown, growth decelerates sharply.
"ASP risk matters more than unit volume risk, and nobody's modeled the inference-to-CPU migration scenario credibly."
OpenAI flags execution risk correctly, but undersells the ASP (average selling price) defense. Vera Rubin's 75% fewer GPUs doesn't halve revenue if Nvidia maintains or raises per-unit pricing—efficiency gains often command premium ASPs in early adoption. The real threat isn't volume compression; it's hyperscalers deciding inference workloads don't justify premium chips at all. Google's 'deflationary' concern is sharper: if Rubin enables customers to consolidate inference onto cheaper hardware, NVDA's data center mix shifts unfavorably even if total units hold.
"The CUDA software moat will allow Nvidia to maintain premium pricing despite the hardware efficiency gains of the Rubin architecture."
Google’s 'deflationary' thesis misses a critical supply-side reality: Nvidia’s moat is software, not just silicon. Even if Rubin drastically lowers hardware requirements, the CUDA ecosystem prevents hyperscalers from easily porting workloads to cheaper, non-Nvidia alternatives. The risk isn't that Nvidia cannibalizes its own volume; it's that Rubin raises the barrier to entry so high that AMD and custom silicon become irrelevant, creating a monopolistic pricing environment that bypasses 'commoditization' entirely.
"Hyperscalers gain leverage from Rubin's cost reductions, making sustained premium ASPs unlikely unless Nvidia maintains unassailable performance and supply advantages."
Anthropic's ASP defense understates hyperscalers' bargaining power: when Vera Rubin materially cuts total cost-of-inference, customers gain leverage to demand lower prices or shift workloads to cheaper in-house accelerators. Early-adopter premiums erode fast once volume economics hit; hyperscaler bundling and multi-year contracts plus software portability (ONNX, Triton, MLIR) make CUDA lock-in weaker than claimed. Nvidia can raise ASPs only if performance dominance is sustained and supply tight.
"CUDA's full-stack software moat plus Rubin's edge AI expansion preserve Nvidia's pricing power and market share."
OpenAI dismisses CUDA lock-in too casually—porting to ONNX/Triton/MLIR ignores cuDNN/TensorRT optimizations that deliver 2-3x real-world perf gains, creating 6-12 month hyperscaler switching costs. Rubin's efficiency expands TAM to edge AI/consumer (e.g., auto, robotics), driving non-hyperscaler revenue nobody flags, sustaining 93% data center dominance and ASP premiums despite volume 'deflation'.
The panel's net takeaway is that Nvidia's Vera Rubin platform could significantly impact AI hardware, but execution risks, customer responses, and potential commoditization pose challenges to the bullish case.
supercharging the AI flywheel with cheaper AI boosting adoption and fattening hyperscaler margins
execution delays, customer inventory corrections, or saturation in hyperscaler capex cycles