$2,304 Invested in Nvidia Today Will Be Worth Triple That Amount in 5 Years
Ahead of Labor Day, you could buy 10 shares of Nvidia (NASDAQ: NVDA) for $2,304. I think those shares could be worth around $6,900 in five years, or about 3 times what the stock trades for today.
Making such a bold claim, of course, requires some proof for the skeptics. First, I’ll get into what makes Nvidia special and why its momentum is likely to continue; then I’ll get into the math showing how the stock could triple by 2030.
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 »
The biggest AI winner
Nvidia has established itself as the biggest artificial intelligence (AI) winner since the technology started to go mainstream. That’s because its graphics processing units (GPUs) are the primary chips powering AI workloads. While there is much more increased competition coming from custom AI chips and the occasional better offerings from Advanced Micro Devices and newer chip upstarts, Nvidia still finds itself in the catbird seat.
The company is the absolute dominant player in the AI training market, and this is unlikely to change. Nvidia has created a wide moat in this area with its CUDA software platform. It developed CUDA to easily program its chips and smartly seeded it among universities and research facilities doing early work on AI. The result is a generation of developers trained on its software, with most early AI code written on its platform and optimized for its chips.
Training is just part of the story, though, as inference is now growing faster and expected to eventually become the larger of the two AI computing markets. CUDA’s moat is not as formidable in this area, but the company has made some good moves to remain a top player in this arena as well. Nvidia smartly “acquired” Groq and its language processing units (LPUs) earlier this year and incorporated them into its CUDA ecosystem. Inference tends to be more memory-bound than compute-bound, and LPUs have lightning-fast SRAM (static random-access memory) directly embedded in them. This reduces latency and makes them ideal for the decode phase of inference. Meanwhile, its GPUs can handle the more compute-heavy pre-fill phase.
This also speaks to Nvidia’s greater strategy. The company is no longer just a GPU maker; it’s become a complete AI infrastructure player. With a world-class networking portfolio, its own central processing units (CPUs), and other chips, the company can now deliver end-to-end complete rack solutions for specific AI tasks, such as training, inference, and agentic AI. At the same time, it’s also acquired AI ecosystem players like SchedMD and Hugging Face (pending) to support open-source AI models and ensure that the broader developer ecosystem relies on open standards that run best on its hardware.
