My prediction that Nvidia (NVDA 1.53%) would beat the S&P 500 (SNPINDEX: ^GSPC) for the third consecutive year came true, as the artificial intelligence (AI) growth stock soared 38.9%, compared with a 16.4% gain for the index.
Here's why Nvidia can beat the S&P 500 again in 2026.
Image source: Getty Images.
Leading the next stage of AI evolution
In late October, Nvidia forecast that it would reach $500 billion in Blackwell and Rubin booked orders, including revenue that would be realized in fiscal 2026. At CES earlier this month, Nvidia said it would exceed its prior $500 billion guidance as orders continue to pour in.
On Jan. 5, Nvidia released a press release announcing six new Rubin chips, including a graphics processing unit (GPU), a central processing unit, and various network, storage, and connectivity products. It also said Rubin is now in full production, with product deliveries coming in the second half of calendar year 2026.
In response, analysts are bumping up their earnings estimates. At the time of this writing, consensus analyst estimates call for fiscal 2026 earnings per share of $4.69 and $7.60 for fiscal 2027. Just six months ago, consensus estimates were $4.29 for fiscal 2026 and $5.76 for fiscal 2027 -- showcasing the impact of pulling forward Rubin earnings. Nvidia's fiscal 2027 ends in late January 2027, so the Rubin windfall is likely to begin in the second half of fiscal 2027.
Despite launching less than two years after Blackwell, Rubin has achieved massive efficiency improvements by packing more transistors per GPU and by leveraging what Nvidia calls "extreme codesign," which is basically another term for vertical integration across the entire technology stack, from rack-scale hardware solutions to software, algorithms, and physical infrastructure.
In its CES presentation, Nvidia showed that its Rubin solution can achieve up to a 90% reduction in inference token cost and a 75% reduction in the number of GPUs needed to train AI models. Tokens are a way to measure the data that AI algorithms process in their workflows. So the more tokens that can be processed thanks to higher compute, the more efficient the operation.
In sum, Nvidia can justify charging high prices for its Rubin products because they should improve performance, and therefore lower data center operating costs for customers.
NASDAQ: NVDA
Key Data Points
Nvidia stays red hot
Nvidia's innovation should allow it to sustain high margin and hold its own against mounting competition. The bigger threat to the long-term investment thesis is demand.
Nvidia's key customers are some of the most profitable companies in the world, with incredibly deep pockets. But eventually, they will need to convert capital expenditures into earnings to justify higher spending.
In the meantime, demand for Rubin indicates that AI spending isn't slowing down, making Nvidia's 39 forward price-to-earnings ratio reasonable. Until demand cools off, there's every reason to believe Nvidia can continue rewarding long-term investors.
