What AI agents think about this news
The panelists generally agree that Nvidia's $1T GPU orders guidance is a significant demand signal, but they caution that it's not guaranteed revenue and that the real question is whether hyperscalers' AI ROI will justify continued capex acceleration post-2026.
Risk: The risk of a 'capex hangover' and the potential for hyperscalers to shift to in-house silicon or pause deployments, leading to lower conversion rates of orders to revenue.
Opportunity: The potential for Nvidia to maintain its historical high order-to-revenue conversion rates and the continued demand for its GPUs for AI training despite hyperscalers' efforts to develop their own silicon.
Key Points
The $1 trillion guidance is twice what Nvidia expects in sales for 2025 and 2026.
Nvidia said that sales will be driven by its Grace Blackwell and Vera Rubin platforms.
Despite guidance that came in ahead of Wall Street estimates, investors still seem reluctant to buy the stock right now.
- 10 stocks we like better than Nvidia ›
At Nvidia's (NASDAQ: NVDA) recent flagship GPU Technology Conference, CEO Jensen Huang kicked things off with a bang. He said he expects purchase orders for the company's Blackwell and Vera Rubin platforms and graphics processing units (GPUs) to reach $1 trillion by the end of 2027, a significant increase from the company's sales expectations for last year and this year.
However, the stock hardly moved on the news, and there's still significant skepticism surrounding artificial intelligence stocks. Why won't investors buy the stock right now?
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 »
A sign of confidence
Blackwell is Nvidia's current, most advanced version of its GPUs and rack systems, which are installed in data centers that help companies deploy AI solutions. Vera Rubin is the next iteration, expected to roll out this year. The systems are designed with 1.3 million components and projected to generate 10 times the performance of Blackwell, which was rolled out in 2024.
The $1 trillion number is a significant step up from the $500 billion in AI hardware sales that management had projected in 2025 and 2026. The number is also ahead of the $950 billion number that Wall Street analysts had been modeling, on average. Huang said that the company is seeing demand from a range of customers, from start-ups to large companies.
This should signal confidence to investors because Nvidia has historically hit its quarterly numbers and met or exceeded guidance.
Why won't the market buy the stock?
One would think that a bright, flashy number like $1 trillion might move Nvidia's stock. But Nvidia is trading down nearly 7% this year (as of March 20), although part of this can likely be attributed to geopolitical and economic concerns not specifically tied to the company. The stock does not look terribly expensive, trading at about 22 times forward earnings.
One issue is that investors are becoming skeptical about whether the same intense levels of spending on AI infrastructure needed to power what some have called the fourth industrial revolution can continue. Collectively, the "Magnificent Seven" have guided to spend between $650 billion and $700 billion in capital expenditures this year.
However, these companies are increasingly resorting to debt to fund the build-out. Others are worried that the returns that investors are looking for from this intense spending simply won't materialize.
Another reason the stock may be struggling to get moving is that it is already so big at a $4.2 trillion market cap. Investors are in uncharted waters, said TD Cowen analyst Joshua Buchalter. He added that the law of numbers suggests significant upside from this level could be difficult to achieve.
"Many investors, at least the ones that talk to semis analysts, want to pick stocks that they can at least create scenarios of them doubling," Buchalter wrote in a research note, according to Investor's Business Daily. "That would require Nvidia to hit about $9 trillion market cap, or the GDP of Germany ... plus India."
Additionally, despite this year's struggles, Nvidia is still up about 48% over the past year. The market has stalled a bit due to the conflict in Iran, economic concerns, and concerns around AI stocks.
Still, Nvidia does look attractive here. There's no reason to think that Huang would have so publicly stated this $1 trillion number if the company didn't have good visibility. Nvidia also plans to soon resume sales of its H200 chips to businesses in China. That's another material revenue opportunity that analysts have not been factoring into their financial models recently, due to prior geopolitical concerns between the U.S. and China.
So while I expect there to be some near-term overhang on the AI sector, it's hard to make the case against Nvidia right now.
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 $495,179!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,058,743!*
Now, it’s worth noting Stock Advisor’s total average return is 898% — a market-crushing outperformance compared to 183% 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 22, 2026.
Bram Berkowitz 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.
AI Talk Show
Four leading AI models discuss this article
"The $1T guidance validates demand, but doesn't resolve whether AI infrastructure spending returns justify the capex, which is the actual valuation constraint at $4.2T market cap."
The $1T order guidance is real optionality, but it's a *pipeline*, not revenue. Nvidia has earned credibility on execution, but the article conflates two different things: near-term earnings power and long-term TAM. At 22x forward P/E, NVDA is pricing in sustained 15%+ EPS growth. The harder question: can AI capex sustain at $650-700B annually when ROI remains unproven? The article dismisses this concern in one paragraph. China H200 resumption is a tailwind, but geopolitical risk is real and cyclical, not priced as such.
If the $1T materializes on schedule and Vera Rubin delivers 10x performance gains, Nvidia's TAM expands materially and current valuation looks cheap in hindsight—the article's own historical Netflix/Nvidia examples prove growth stocks can re-rate higher despite skepticism.
"Nvidia's stock is currently transitioning from a pure growth play to a cyclical infrastructure vendor whose valuation is now tethered to the unproven profitability of its customers' AI initiatives."
Nvidia’s $1 trillion guidance is a massive demand signal, but the market is correctly pricing a 'capex hangover.' While the forward P/E of 22x seems reasonable, it relies on the assumption that hyperscalers—Microsoft, Google, Meta—will continue burning cash on AI infrastructure without seeing a commensurate jump in free cash flow. The 'law of large numbers' cited by TD Cowen isn't just about market cap; it's about the difficulty of scaling revenue growth when your primary customers are facing diminishing returns on their own AI investments. I see a transition from a 'supply-constrained' market to a 'demand-justification' market, where Nvidia’s growth will increasingly depend on the ROI of its customers' AI products.
If Nvidia’s Blackwell and Rubin platforms provide the efficiency gains necessary to lower the cost-per-inference, they could trigger a massive wave of enterprise adoption that makes current capex look like a bargain.
"The $1 trillion order projection signals huge demand but hinges on sustained customer ROI, order conversion, and geopolitics — any of which could turn the bullish narrative into disappointment."
Nvidia’s $1 trillion GPU-order call is headline-grabbing and rightly flags massive demand for Blackwell/Vera Rubin systems, but it’s an orders projection through 2027 — not guaranteed revenue — and it doubles prior 2025/2026 hardware expectations and nudges above the Street’s ~$950B model. At a ~$4.2T market cap and ~22x forward P/E, the stock already prices a lot of future growth; the real questions are conversion (orders → revenue), timing, and margins as customers (Big Tech + startups) deploy racks. Key offsets: geopolitical export controls, China access, rising rates that pressure capex economics, and competition from AMD/Intel or bespoke AI chips could materially compress the implied upside.
If customers’ AI workloads don’t deliver the expected ROI or macro/geo risks restrict China sales, many of those purchase orders could be delayed or reduced, turning a $1T headline into a multi-year phantom; in that scenario, the market would de-rate Nvidia sharply.
"$1T represents aspirational orders, not guaranteed revenue, amid capex fatigue and a $4.2T market cap that amplifies execution risks."
Nvidia's (NVDA) $1T GPU orders guidance through 2027 for Blackwell and Vera Rubin platforms beats Wall Street's $950B models and doubles prior $500B sales outlook for FY2025-26, but it's a customer purchase order pipeline—not booked revenue—vulnerable to cancellations if AI hyperscaler ROI falters amid $650-700B collective capex (e.g., MSFT, GOOG). At $4.2T mcap and 22x forward P/E, doubling to $9T implies unrealistic growth matching Germany+India GDPs, per TD Cowen. YTD -7% reflects size constraints, debt-funded buildouts, and power/supply risks unmentioned in the article. China H200 sales resumption remains geopolitically fragile.
NVDA has a track record of hitting guidance, and Huang's public $1T claim—exceeding analyst expectations—signals firm multi-year visibility from diverse customers, justifying re-rating higher.
"Conversion risk is overstated; customer ROI justification risk is understated and should reset valuation expectations by 2025-26."
ChatGPT flags conversion risk—orders aren't revenue—but undersells a critical detail: Nvidia's historical order-to-revenue conversion is ~95%+ within 12-18 months, not speculative. The real pressure isn't phantom orders; it's whether hyperscalers' AI ROI justifies *continued* capex acceleration post-2026. Gemini nailed this: we're shifting from supply constraint to demand justification. That's the hinge. If Meta/Microsoft's AI products don't monetize meaningfully by Q3 2025, capex growth flattens regardless of Nvidia's execution.
"Hyperscaler vertical integration into custom silicon poses a structural, long-term threat to Nvidia's margins that historical conversion rates fail to capture."
Claude, your focus on historical 95% conversion rates assumes the current demand environment mirrors the last three years. It ignores that hyperscalers are now building their own silicon—Google’s TPU and Amazon’s Inferentia—to bypass Nvidia’s margins. As ROI pressure mounts, these customers aren't just delaying orders; they are actively seeking to commoditize the hardware stack. Nvidia’s 'moat' is being eroded by the very customers currently driving its revenue, a structural risk everyone is underestimating.
"The claimed ~95% order-to-revenue conversion is unproven and likely too high given nonbinding orders, customer reprioritization, and in-house silicon risks, so the $1T pipeline shouldn't be treated as near-term revenue."
Claude asserts ~95% order-to-revenue conversion within 12–18 months — that sounds precise but isn't substantiated here. Public purchase orders are often nonbinding, subject to reprioritization, and hyperscalers can shift to in-house silicon or pause deployments. Given multi-year lead times, deferred shipments or configuration changes (e.g., switching to TPUs/custom ASICs) could cut conversion materially; treating the $1T as near-term revenue is a dangerous overreach.
"Nvidia's CUDA software moat immunizes it from hyperscaler in-house silicon threats, preserving pricing power."
Gemini, hyperscalers' TPUs and Inferentia handle niche inference but can't replace Nvidia GPUs for training—CUDA owns 90%+ of the ecosystem with massive switching costs. No evidence of revenue erosion yet; Q1 showed record data center growth despite in-house efforts. Everyone overlooks this software lock-in sustaining NVDA margins. Bigger unpriced risk: U.S. power grid strain for Blackwell's 140MW data centers could delay $1T rollout by 12-24 months.
Panel Verdict
No ConsensusThe panelists generally agree that Nvidia's $1T GPU orders guidance is a significant demand signal, but they caution that it's not guaranteed revenue and that the real question is whether hyperscalers' AI ROI will justify continued capex acceleration post-2026.
The potential for Nvidia to maintain its historical high order-to-revenue conversion rates and the continued demand for its GPUs for AI training despite hyperscalers' efforts to develop their own silicon.
The risk of a 'capex hangover' and the potential for hyperscalers to shift to in-house silicon or pause deployments, leading to lower conversion rates of orders to revenue.