Led by CEO and co-founder Jensen Huang, Nvidia (NVDA 1.97%) has established itself as the top chipmaker in AI, and it does not plan on giving up its throne anytime soon. Much of the company’s success can be directly tied to Huang’s instinctive talent for predicting where the tech world is headed well in advance. That’s why the stock is a buy.
Nvidia was founded in 1993, and its invention of the graphics processing unit (GPU) in 1999 helped fuel the video game market by speeding up graphics rendering and allowing for major leaps forward in computer graphics. The video game market was big at the time, but Huang’s more important strategic move was to have Nvidia create its CUDA software platform, which makes its chips programmable for other tasks.
Today’s Change
(-1.97%) $-4.09
Current Price
$203.31
Key Data Points
Market Cap
Day’s Range
$197.97 – $206.64
52wk Range
$164.07 – $236.54
Volume
4.6M
Avg Vol
156.7M
Gross Margin
74.15%
Dividend Yield
0.14%
A history of forward-looking moves
While the full value of that strategy took many years to manifest, Nvidia wisely seeded CUDA into universities and research labs that were doing early work on AI. The result was that most foundational AI code was written on CUDA for Nvidia’s GPUs, which is why the company enjoys a wide moat in AI model training today.
Huang did not stop there, though. In 2020, Nvidia acquired networking company Mellanox. It was a company with technology ahead of its time, but Huang again saw where the market was heading. Today, Nvidia’s networking portfolio is the fastest-growing part of its business and a key part of its transformation from a GPU specialist into a complete AI infrastructure player.
Huang also anticipated the shift toward inference and agentic AI, and took steps to ensure Nvidia would be a big player in these markets. The company has developed its own ARM-based central processing units (CPUs), as CPUs will play an important role in managing AI agents. The GPU-to-CPU ratio in AI data centers built when workloads were primarily driven by training was 8 to 1. As cloud companies build out infrastructure for agentic AI, the prediction is that the ratio could shift to 1 to 1. With that in mind, Nvidia has projected that the data center CPU market could reach a value of $200 billion in the next few years.
Image source: Nvidia.
Nvidia also acquired the assets and key personnel of Groq, including its language processing units (LPUs), which it has since incorporated into the CUDA ecosystem. These chips will help with servers designed specifically for inference, a market that’s eventually expected to grow to a much larger size than AI model training.
The company’s unique server offering will use both GPUs and LPUs, with GPUs handling the prefill phase of understanding users’ prompts and LPUs dealing with the decode phase of giving quicker responses. This could be the next big growth driver for the company.
Overall, Nvidia is an attractively priced stock. It’s trading at just 16 times analysts’ earnings estimates for its fiscal 2028 (which ends in January 2028), and its top and bottom lines are growing rapidly. However, the biggest reason to own this AI stock for the long term is that Huang has proven to be a visionary who can position Nvidia for what’s next.
