Nvidia's Rubin platform is arriving just as the AI race is pivoting from training to inference.
Customers have already lined up for Rubin, and Nvidia's sales could double by the end of next year.
But investors will want to consider the stock's valuation and risks before hitting the buy button.
Nvidia (NASDAQ: NVDA) has dominated the artificial intelligence (AI) data center market with its GPU chips for years now. The AI era began with the Hopper microarchitecture, continued with Blackwell, and now looks to Rubin, Nvidia's most ambitious chip platform yet.
The stakes are also arguably the highest that they've ever been. Not only has the market begun questioning AI hyperscalers for pouring hundreds of billions of dollars into data centers, but AI technology itself is evolving rapidly, creating a moving target for Nvidia and others.
Where to invest $1,000 right now? Our analyst team just revealed what they believe are the 10 best stocks to buy right now, when you join Stock Advisor. See the stocks »
Rubin is in full production to begin shipping later this year. Is now the time to buy Nvidia stock, or is the AI chip king about to lose its crown?
Image source: Nvidia.
Nvidia specializes in GPU chips that excel at high-performance computing, such as training AI models. That's still important as OpenAI and other AI companies develop more advanced models. But now, AI is quickly shifting toward inference. Deloitte Global predicted that compute for inference will overtake training this year.
OK, what does that mean?
Think of AI chips as engines in a car. A GPU's raw power would be excellent for going very fast in a straight line, like in a drag race. But inference is a different type of compute, a different type of race. If it were a car, it would be the winding mountain roads that would require a more refined engine. In AI, inference compute involves strong, sustained reasoning capabilities for applications such as AI agents. In these situations, efficiency and speed trump raw maximum power.
Rubin isn't a single GPU; it's six components, including CPUs and GPUs, networking switches and interfaces, that all work together as an AI supercomputer. Nvidia designed Rubin with inference in mind. The Vera Rubin NVL72 server rack-mounted superchip performs inference computing at 10% of the cost per million tokens compared to Nvidia's current flagship Blackwell GB200 NVL72.
AI hyperscalers are getting pressure to begin realizing a return on their AI spending, and the drastic improvements in inference efficiency will likely be a strong selling point for Rubin. Nvidia announced back in November that it already had a $500 billion order backlog through 2026, and that is likely to grow further once it reports its latest earnings on Feb. 25. Just about every AI hyperscaler plans to spend even more on AI in 2026.

Data by YCharts.
Wall Street analysts certainly expect big things ahead for Nvidia. Estimates anticipate the company's trailing-12-month sales of $187 billion increasing to approximately $327 billion this fiscal year, and $419 billion the following year. On top of that, Nvidia has beaten Wall Street's quarterly revenue estimates for the past three years without fail, so it wouldn't be a surprise if the actual numbers were even higher.
Nvidia stock trades at nearly 25 times sales. If the company performs as expected, the stock gets much cheaper fairly quickly. Shares trade at just 11 times next year's revenue estimate. It's worth noting that Nvidia depends on a very concentrated customer base, and one or more hyperscalers backing off on their AI spending could quickly turn those estimates on their head.
Even though Nvidia still holds the lion's share of the installed base of AI data center chips, it must continue to fend off competition, especially if customers place greater emphasis on cost savings. Broadcom has already had great success partnering with Alphabet on Tensor Processing Units (TPUs) for its AI models.
Nvidia will likely remain a long-term AI winner due to its entrenched leadership and the growth opportunities ahead for AI as a whole. However, investors may want to hold off, or at least keep buying to a nibble, until Nvidia updates shareholders on its latest outlook following earnings.
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 $424,262!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,163,635!*
Now, it’s worth noting Stock Advisor’s total average return is 904% — a market-crushing outperformance compared to 194% 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.
See the 10 stocks »
*Stock Advisor returns as of February 23, 2026.
Justin Pope has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet and Nvidia. The Motley Fool recommends Broadcom. The Motley Fool has a disclosure policy.