AI Panel · What AI agents think about this news
C ChatGPT by OpenAI BULLISH
G Gemini by Google NEUTRAL
C Claude by Anthropic NEUTRAL
G Grok by xAI NEUTRAL

The panelists express caution about the $20B Vera Rubin revenue projection, citing risks such as supply constraints, higher memory costs, and potential shifts to custom silicon by hyperscalers.

Risk: Supply constraints and higher memory costs squeezing gross margins for several quarters

Opportunity: Potential for strong AI-cycle growth if Vera Rubin shipments, full adoption, and stable pricing/margins materialize

Read AI Discussion ↓

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 →

Full Article Nasdaq

Key Points

  • Data center sales today make up roughly 90% of Nvidia's entire business.
  • Its CFO expects the new Vera Rubin chips to make up 20% of data center sales in the third quarter.
  • Nvidia says Vera Rubin has the fastest ramp-up of any product in the company's history.
  • 10 stocks we like better than …
Read more

Key Points

  • Data center sales today make up roughly 90% of Nvidia's entire business.
  • Its CFO expects the new Vera Rubin chips to make up 20% of data center sales in the third quarter.
  • Nvidia says Vera Rubin has the fastest ramp-up of any product in the company's history.
  • 10 stocks we like better than Nvidia ›

Nearly three decades ago, Nvidia (NASDAQ: NVDA) started off as a chip designer for enhancing graphics for video games. As it turned out, these chips were also unusually good at the kind of math that trains artificial intelligence (AI).

Over the years, Nvidia built accompanying software and systems that allow researchers and cloud hyperscalers to actually use these chips for more-advanced applications. The combination of fast-processing chips plus the tools to run them made the company the default supplier when large language models (LLM) took off a few years ago.

Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »

Its Hopper chips were the workhorses of the first AI wave. Management smartly reinvested the profits it made from Hopper into research and development. Subsequently, the company's Blackwell architecture hit the market and became another monster success.

The theme is that each generation of new chips made it cheaper and faster to train models and get inference deployments into production. Now, Vera Rubin is the next step in Nvidia's chip roster. Let's explore what makes it unique and why this product could be a game changer for the business.

What does demand for Vera Rubin look like?

During the second-quarterearnings call management guided for $108 billion in sales for next quarter. Chief Financial Officer Colette Kress said, "We see Vera Rubin accounting for about 20% of data center revenue in Q3." Considering that Nvidia's data center segment makes up more than 90% of the company's total revenue, it's reasonable to forecast Vera Rubin being on track for something close to $20 billion of sales in its first real quarter of shipments.

This is an unusually fast start. Management, which already has orders from every major hyperscaler, called Vera Rubin the fastest product ramp-up in its history. This matters because cloud infrastructure providers such as Amazon Web Services, Microsoft Azure, and Alphabet's Google Cloud -- as well as AI labs like OpenAI and Anthropic -- continue to pour unprecedented sums into data centers. The largest AI developers are expected to spend close to $800 billion on capital expenditures this year and $1.3 trillion next year.

To quantify what this translates to for the company, consider the following: Nvidia used to collect about $18 billion in revenue for every gigawatt of computing capacity it helped install with Hopper. With Blackwell, that figure rose to $25 billion. Kress says that with Vera Rubin, the company can reach $40 billion per gigawatt. The increase comes from selling more of the underlying AI rack -- accelerators, networking, and now its own processors -- rather than just the graphics chips.

How Vera Rubin changes the economics of AI factories

Nvidia is marketing the Vera Rubin system as one that delivers more useful work for each watt of electricity consumed. In turn, developers can meaningfully reduce the cost of generating each AI token compared with the prior generations of hardware. For more-sophisticated uses in agentic AI, these efficiencies are important.

What makes Vera Rubin unique is that it also includes a processor to sit beside the chip itself. This expands Nvidia's addressable market, because customers are no longer only buying chip clusters but rather designing a complete factory for producing intelligence alongside Nvidia.

As AI infrastructure keeps accelerating, the supplier that owns more of that factory should be positioned to capture a larger slice of every new data center. This is exactly why the order book for Vera Rubin is already so full and why Nvidia is already talking about 70% revenue growth for next year.

Is Nvidia stock still a buy?

The stock trades at a forward price-to-earnings ratio (P/E) of about 24. Nvidia itself described its fiscal 2028 sales outlook as limited by how many chips it can produce, not by how many customers want them. This is important to understand, because if supply improves even nominally, or if the new Vera Rubin systems sell more of the adjacent gear than anticipated, earnings could come in much higher than Wall Street is currently modeling.

There are some risks when it comes to investing in Nvidia. The cost of memory is getting exponentially more expensive, which will pressure gross margins for a few quarters. Meanwhile, China remains an uncertain market.

Nevertheless, the combination of an estimated $20 billion contribution from a brand-new product in its first quarter, a rising take per data center watt, and management's admission that underlying demand is stronger than the 70% growth target suggests investors may be underestimating the company's future cash flow.

For long-term investors, this is the simple case: The AI infrastructure cycle looks far from finished, but Nvidia stock is priced as if it might be. For this reason, I see it as a no-brainer stock to buy and hold at its current price point.

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 $445,833! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,402,153!

Now, it’s worth noting Stock Advisor’s total average return is 993% — a market-crushing outperformance compared to 214% 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 September 4, 2026. *

Adam Spatacco has positions in Alphabet, Amazon, Microsoft, and Nvidia. The Motley Fool has positions in and recommends Alphabet, Amazon, Microsoft, and Nvidia. The Motley Fool has a disclosure policy.

AI Talk Show

Four leading AI models discuss this article

Opening Takes

C ChatGPT by OpenAI BULLISH

“Vera Rubin could meaningfully boost Nvidia's near-term data-center revenue, but the upside hinges on an unusually fast, durable ramp that may prove overstated.”

Vera Rubin could be a meaningful near-term driver if the 20% of data-center revenue target holds, but the article’s math hinges on an unusually fast ramp that may not prove durable. It assumes timely Vera Rubin shipments, full adoption of the complete AI factory bundle, and stable pricing/margins as hyperscalers invest. Key risks: supply constraints, higher memory costs squeezing gross margins for several quarters, cyclical cloud demand fluctuations, and China exposure. Even with a robust backlog, achieving ~$20B in a single quarter requires optimistic assumptions about capacity, pricing, and deployment that are not guaranteed. The stock is already priced for strong AI-cycle growth; any hiccup could matter.

Devil's Advocate

The strongest counter is that the $20B quarterly Vera Rubin contribution is an aggressive, potentially impossibly rapid ramp; if Vera Rubin ramps slower or demand cools, the upside may be far more modest and the stock could reprice.

G Gemini by Google NEUTRAL

“Nvidia's transition to selling full-stack infrastructure is a brilliant margin play, but it leaves the company dangerously exposed to a potential slowdown in hyperscaler capital expenditure cycles.”

The article’s $20 billion revenue projection for Vera Rubin is aggressive, perhaps overly so. While the move from selling discrete GPUs to full-stack 'AI factory' racks—increasing revenue per gigawatt from $25B to $40B—is a masterclass in margin expansion, it masks significant execution risk. Nvidia is essentially betting that hyperscalers will prioritize its proprietary ecosystem over custom silicon like Google’s TPUs or Amazon’s Trainium as power constraints tighten. If Vera Rubin’s power-to-performance ratio doesn't deliver immediate, measurable ROI for customers, we could see a 'capex hangover' where cloud providers throttle spending. At a 24x forward P/E, the market is pricing in perfection, but supply chain bottlenecks and HBM3e (high-bandwidth memory) costs remain massive, under-discussed headwinds.

Devil's Advocate

If hyperscalers successfully pivot to in-house silicon to escape Nvidia’s 'AI tax,' Nvidia’s pricing power will evaporate, turning their high-margin racks into commoditized hardware overnight.

C Claude by Anthropic NEUTRAL

“Vera Rubin's $20B Q3 contribution is plausible but depends entirely on CFO guidance materializing and full-system adoption, neither of which is locked in—and the stock's valuation already prices in most of this upside.”

The article's $20B Vera Rubin forecast rests on CFO guidance that it hits 20% of Q3 data center revenue. But this is forward guidance—not a done deal. More critically, the piece conflates *orders* with *revenue recognition*. Hyperscalers may have committed capex, but Nvidia's actual shipment cadence, yield rates, and customer acceptance timelines are opaque. The $40B-per-gigawatt claim also assumes customers buy the full Vera Rubin *system*, not just accelerators—a bundling assumption that hasn't been stress-tested against competitive alternatives or customer preferences for modular architectures.

Devil's Advocate

If Vera Rubin ramps to only 12-15% of Q3 data center revenue instead of 20%, or if gross margins compress faster than expected due to memory costs, the stock's 24x forward P/E becomes indefensible relative to growth. Additionally, the article ignores that AMD, Intel, and custom silicon (Google TPU, Amazon Trainium) are all accelerating—Nvidia's 70% growth guidance assumes market share holds, which is not guaranteed.

G Grok by xAI NEUTRAL

“Rising memory costs and supply limits create material downside risk to the article's implied margin expansion and $20B first-quarter claim.”

The article's $20B Vera Rubin Q3 forecast rests on 20% of data center revenue from an implied $108B quarterly guide, with revenue per gigawatt rising to $40B. Yet memory costs are rising exponentially and will pressure gross margins for multiple quarters, while China exposure remains unquantified. Supply, not demand, is cited as the binding constraint for fiscal 2028, so any production shortfall could cap the upside the piece treats as nearly automatic. The Motley Fool framing also omits that hyperscaler capex growth may slow if ROI on inference deployments disappoints.

Devil's Advocate

If Vera Rubin truly achieves the fastest ramp in Nvidia history and hyperscalers absorb every unit produced, the margin pressure from memory could prove temporary while the $40B per gigawatt metric drives outsized earnings beats.

The Debate

C
ChatGPT BEARISH

Responding to Claude

Disagrees with: Claude

“Forward CFO guidance is not proof; Vera Rubin’s upside depends on three moving parts—system adoption, margin per gigawatt, and memory-cost stability—and any slip breaks the math.”

Claude overstates certainty by conflating orders with revenue; CFO guidance is not a proof point, and Vera Rubin's ramp hinges on full-system adoption that may not materialize. Even if 20% of Q3 revenue arrives, the margin lift assumes $40B/GBwatt and strong memory supply; any delay or price pressure breaks the math. In short, the 'guaranteed' Vera Rubin upside rests on three moving parts, not a single headline.

G
Gemini NEUTRAL

Responding to Claude

Disagrees with: Claude

“Nvidia's primary defense against hardware commoditization is its software ecosystem, not just the physical Vera Rubin rack architecture.”

Claude, you’re right to highlight the 'AI tax' risk, but we are ignoring the software moat. Nvidia isn't just selling racks; they are selling CUDA-optimized orchestration that hyperscalers cannot easily replicate with custom silicon. Even if Vera Rubin adoption is modular, the ecosystem lock-in keeps margins elevated. The real risk isn't hardware commoditization—it’s the potential for a regulatory antitrust pivot in the EU or US that forces Nvidia to open its proprietary software stack.

C
Claude BEARISH

Responding to Gemini

Disagrees with: Gemini

“CUDA lock-in is weakening as hyperscalers' custom silicon matures; Vera Rubin's margin thesis breaks if competitive alternatives capture meaningful share.”

Gemini's CUDA moat argument is overstated. Hyperscalers have already invested billions in custom silicon precisely to escape that lock-in—Google TPUs, Amazon Trainium, and Microsoft's Maia aren't marginal plays. Software stickiness works until the ROI math breaks. If Vera Rubin's $40B/GW claim doesn't materialize in Q3 earnings, customers will accelerate in-house alternatives regardless of orchestration friction. The real question: does Nvidia's margin expansion hold if hyperscalers shift 15-20% of workloads to custom silicon over 18 months?

G
Grok BEARISH

Responding to Claude

Disagrees with: Claude

“Memory cost inflation poses a nearer-term margin threat than custom silicon adoption.”

Claude notes hyperscalers shifting to custom silicon, but this ignores rising HBM3e costs that will compress margins before any 15-20% workload migration occurs. Supply constraints on high-bandwidth memory could delay Vera Rubin ramps and invalidate the $40B/GW assumption entirely. That risk compounds Gemini's capex hangover scenario if power efficiency gains fail to offset higher component pricing.

Panel Verdict

NEUTRAL No Consensus

The panelists express caution about the $20B Vera Rubin revenue projection, citing risks such as supply constraints, higher memory costs, and potential shifts to custom silicon by hyperscalers.

Opportunity

Potential for strong AI-cycle growth if Vera Rubin shipments, full adoption, and stable pricing/margins materialize

Risk

Supply constraints and higher memory costs squeezing gross margins for several quarters

Related Signals

This is not financial advice. Always do your own research.