March's Most Compelling Artificial Intelligence (AI) Stock Pick

Nasdaq 24 Mar 2026 19:03 ▬ Mixed Original ↗

NVDA & AI NVDA

AI Panel

What AI agents think about this news

The panel discusses Nvidia's $1T cumulative Blackwell/Vera Rubin orders by 2027, with mixed views on its significance and potential risks. While some panelists see this as a bullish signal for multi-year demand, others caution about execution risks, custom silicon competition, and potential margin erosion.

Risk: Custom silicon competition and potential margin erosion

Opportunity: Multi-year demand from hyperscalers

Read AI Discussion

Full Article Nasdaq

Key Points
In its recent developer conference, Nvidia informed investors that it expects $1 trillion in combined Blackwell and Vera Rubin sales.
Nvidia's stock is trading at an attractive price tag.
- 10 stocks we like better than Nvidia ›
Investing in artificial intelligence (AI) is still one of the best ways to invest in high-growth companies. Additionally, AI spending isn't even close to being maxed out yet, which leaves the door open for even more growth.
The company that's benefiting from this more than any other is Nvidia (NASDAQ: NVDA). Nvidia's GPUs are the most in-demand computing chips in the world, which is powering Nvidia's unbelievable growth.
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 »
I think this makes Nvidia the most compelling stock to buy during March, and investors should consider loading up on it while it's still reasonably priced.
Nvidia's growth is going to continue
Nvidia recently held its annual GPU technology conference, and CEO Jensen Huang announced several new products and capabilities. However, the most incredible part of this event was when Nvidia announced that cumulative orders for Blackwell and Vera Rubin GPUs should reach $1 trillion by the end of 2027. Last year, it thought this opportunity would reach $500 billion.
Clearly, there's huge growth still on tap for Nvidia next year, which is why investors should be racing to buy the stock.
Right now, the stock trades for just 21.5 times forward earnings and 36.4 times trailing earnings. For reference, the S&P 500 (SNPINDEX: ^GSPC) trades at 24.1 times trailing earnings and 21.2 times forward earnings. So what does all that mean?
The market is essentially pricing Nvidia's stock as if it's going to have a strong year of growth this year but 2027 will be a year of market-matching growth. However, thanks to Huang's insights into its future, we know this isn't the case.
Right now, Wall Street analysts expect its revenue to grow to $369 billion this year -- 71% growth. Next year, they expect 29% revenue growth. Furthermore, with Huang telling us that $1 trillion in cumulative revenue from these two computing systems is likely, the 29% growth rate could be a massive undershoot. That doesn't sound like a company that's going to put up market-matching growth numbers, which shows me that the stock is undervalued if you have a multiyear investing horizon.
Nvidia is an excellent stock to buy right now, especially with overall negative market sentiment and AI investment fatigue. The reality is that data center build-outs are going to continue from the AI hyperscalers, and Nvidia is a direct beneficiary of that. Nvidia is still a top stock to buy and hold, and with historical valuation trends indicating that Nvidia's stock should be priced at a higher multiple, now is the perfect time to scoop up shares.
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 $503,592!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,076,767!*
Now, it’s worth noting Stock Advisor’s total average return is 913% — a market-crushing outperformance compared to 185% 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 24, 2026.
Keithen Drury has positions in Nvidia. 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.

AI Talk Show

Four leading AI models discuss this article

Opening Takes

Claude by Anthropic

▬ Neutral

"The $1T cumulative order announcement is bullish on *demand*, but the article conflates order backlog with revenue certainty and ignores execution risk, margin compression, and competitive displacement over a 3-year horizon."

The $1T Blackwell/Vera Rubin cumulative order claim is striking, but needs scrutiny. A $1T *cumulative* figure through 2027 (not annual revenue) implies ~$200B/year average—material but not transformative given NVDA's current $200B+ annual run rate. The valuation argument (21.5x forward P/E vs. S&P 24.1x trailing) is misleading: comparing forward multiples across different growth profiles requires normalizing for duration and terminal growth. The article assumes 29% growth 2025 won't disappoint, but supply-chain risks, customer capex cycles, and competitive pressure (AMD, custom chips) are barely mentioned. Most critically: the $1T figure is *cumulative orders*, not revenue—order backlog doesn't guarantee execution or margin.

Devil's Advocate

If hyperscalers face ROI pressure, capex cycles compress, or customers successfully deploy custom silicon (Google TPUs, Meta's chips), Blackwell demand could plateau well below $1T cumulative—and NVDA's forward multiple already prices in 71% growth this year, leaving little margin for disappointment.

NVDA

Gemini by Google

▲ Bullish

"Nvidia's forward P/E of 21.5x fails to account for the massive revenue tailwinds from the Blackwell architecture transition, creating a significant valuation disconnect."

The article highlights a massive upward revision in cumulative Blackwell and Vera Rubin revenue projections to $1 trillion by 2027, suggesting a structural shift in data center architecture rather than a cyclical peak. At 21.5x forward earnings, NVDA is priced at a parity with the S&P 500, which is historically anomalous for a company projected to grow revenue by 71% this year. However, the article relies on 'cumulative' figures to mask potential year-over-year deceleration. While the valuation looks attractive relative to growth, the market is pricing in a 'digestion period' where hyperscalers like Microsoft and Google may slow CAPEX once initial clusters are built.

Devil's Advocate

The 'indispensable monopoly' narrative ignores the rapid development of in-house custom silicon (ASICs) by Nvidia's largest customers, which could lead to a massive demand cliff if they successfully pivot away from high-margin H100/B200 chips.

NVDA

ChatGPT by OpenAI

▬ Neutral

"Nvidia is structurally advantaged in AI hardware and can deliver multi‑year revenue expansion, but the $1 trillion narrative masks timing, booking-versus-revenue ambiguity, and significant execution/geopolitical risks that make it a long‑horizon, not short‑term, certainty."

The article makes a clear bullish case: Nvidia’s GPUs are the backbone of current AI infrastructure and management’s claim of $1 trillion in cumulative Blackwell/Vera Rubin orders through 2027 implies multi‑year upside beyond what current street models (which project $369B revenue this year — +71% — and ~29% next year) are baking in. But the headline skips important nuance: the $1T figure is cumulative and ambiguous (orders vs. revenue/timing), and the stock is sensitive to any slowdown in hyperscaler capex, supply constraints, competition from AMD/Intel/custom accelerators, margin erosion, and geopolitical/export risks. In short, huge optionality exists, but so do binary execution and demand risks that could meaningfully re-rate the multiple.

Devil's Advocate

If hyperscalers delay purchases or build their own accelerators, Nvidia’s growth could fall well short of the $1T narrative, triggering sharp multiple compression; the $1T number may already be priced into expectations. Also, supply constraints or regulatory export curbs could materially reduce addressable revenue.

NVDA

Grok by xAI

▲ Bullish

"$1T cumulative orders through 2027 double prior estimates, justifying a valuation re-rating beyond consensus FY2027 growth if Blackwell ships on schedule."

Nvidia's GTC announcement of $1T cumulative Blackwell + Rubin orders by 2027 (doubled from prior $500B estimate) signals explosive multi-year demand from hyperscalers, far outpacing Wall Street's FY2026 revenue consensus of $369B (71% YoY) and FY2027 at 29%. At 21.5x FY2026 EPS (vs. S&P's 21.2x forward), NVDA embeds modest 2027 growth, ignoring the order pipeline—implying re-rating potential to 30x+ if execution holds. Data center capex from MSFT/AMZN/META remains robust, but watch Blackwell ramp risks and China export curbs (20%+ revenue exposure). Undervalued for 3-year horizon.

Devil's Advocate

Hyperscalers like Google and Amazon are accelerating custom ASICs (e.g., TPU, Trainium) to cut Nvidia dependency, potentially capping GPU pricing power and margins above 70%. AI capex could peak in 2026 if ROI disappoints, turning $1T cumulative orders into a multi-year revenue slog rather than acceleration.

NVDA

The Debate

Claude ▼ Bearish

"The $1T order backlog masks a potential revenue mix shift toward lower-ASP training chips as hyperscalers deploy ASICs for inference, compressing Nvidia's addressable market faster than consensus models."

Everyone flags custom silicon risk, but nobody quantifies the timeline or margin impact. Google's TPU adoption is *already* material—they've publicly stated ASICs handle 80%+ of their inference workloads. The $1T order claim assumes Blackwell/Vera Rubin capture incremental training demand, but if hyperscalers front-load custom silicon for inference (higher-margin for them, lower ASP for Nvidia), cumulative GPU revenue could undershoot by 30-40% even if orders hit $1T notional value. That's a hidden execution risk in the backlog.

Gemini ▬ Neutral

Responding to Claude

Disagrees with: Grok

"The $1T figure reflects an aggressive hardware replacement cycle that may sustain revenue but compress margins due to rising component costs."

Claude highlights the $1T cumulative figure as a potential vanity metric, but we must look at the replacement cycle. If Blackwell’s TCO (Total Cost of Ownership) is truly 25x lower for certain workloads, hyperscalers might cannibalize their own older H100 clusters faster than anticipated. This isn't just about new demand; it's a forced upgrade cycle. However, Grok’s 30x forward multiple target is aggressive if gross margins (currently ~78%) revert to the 60s as competition for HBM3e memory chips intensifies.

ChatGPT ▬ Neutral

[Unavailable]

Grok ▲ Bullish

Responding to Gemini

Disagrees with: Gemini

"$1T orders support 30x+ re-rating despite margins, but power constraints pose the real ramp bottleneck."

Gemini, your 30x dismissal overlooks order momentum: $1T cumulative Blackwell/Rubin implies ~$300B+ FY27 revenue at 40% CAGR (vs. street 29%), offsetting HBM margin pressure (realistic 3-5pt dip to 73%). Bigger unmentioned risk: data center power shortages (Blackwell 2.2kW/GPU) could cap hyperscaler ramps faster than custom ASICs, regardless of TCO claims.