Prediction: Vera Rubin Is About to Become a $20 Billion Windfall for Nvidia by Next Quarter

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.

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.

Image source: Nvidia.

What does demand for Vera Rubin look like?

During the second-quarter earnings 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.

NVDA PE Ratio (Forward) data by YCharts.

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.

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.