NVIDIA (NASDAQ:NVDA) stock doubled this year, defying critics who claimed that it would run out of steam after more than tripling last year. Many investors are likely wondering if they should simply buy NVIDIA regardless of its past performance, or if they should wait for a pullback to start a position. Let's examine the bull and bear cases to decide.

What the bulls think of NVIDIA

NVIDIA's core businesses are firing on all cylinders. Last quarter, its gaming GPU revenue rose 25% annually to $1.56 billion, thanks to robust demand for its new GeForce cards. Its data center GPU revenue more than doubled to $501 million, as an increasing number of enterprise customers installed its higher-end GPUs in their data centers for machine learning purposes.

NVIDIA CEO Jensen Huang.

NVIDIA CEO Jensen Huang. Image source: NVIDIA.

Professional visualization revenue rose 15% to $239 million, automotive revenue grew 13% to $144 million, and OEM/IP revenue grew 3% to $191 million.

Sales of NVIDIA's Tegra CPUs -- which power automotive infotainment and navigation systems, Shield gaming devices, and the Nintendo Switch -- rose 74% annually to $419 million. GPU sales rose 31% to $2.22 billion.

The bulls believe that NVIDIA's sales will keep rising across all those categories, as its newer gaming GPUs hold AMD (NASDAQ:AMD) at bay, more enterprise customers buy data center GPUs, and sales of Tegra CPUs continue rising on demand from automakers and Nintendo. They also believe that the growing interest in cryptocurrency mining will lift sales of its GPUs.

That's why analysts expect NVIDIA's revenue and non-GAAP earnings to respectively rise 37% and 63% this year. The bulls will claim those growth rates justify its trailing P/E of 53, which is nearly double the industry average of 27 for semiconductor makers.

What the bears think of NVIDIA

However, the bears will point out that NVIDIA's growth is decelerating, with analysts expecting its sales and earnings to grow just 16% and 12%, respectively, next year. They'll also say that those estimates don't justify NVIDIA's forward P/E of 43.

NVIDIA's growth will likely slow down due to competitive pressures on multiple fronts. AMD's new Vega-based Radeon cards aren't much of a threat to NVIDIA's current-gen Pascal cards or next-gen Volta cards, but AMD could pull a rabbit out of its hat with upgraded Vega chips next year.

Intel (NASDAQ:INTC) also recently partnered with AMD to produce a new multichip processor package with integrated Radeon graphics, which could reduce the need for NVIDIA's discrete GPUs. Intel also launched a new Core and Visual Computing group to create discrete GPUs and hired Raja Koduri, AMD's former senior president and chief architect of Radeon, as its new leader.

Servers in a data center.

Image source: Getty Images.

Intel is also gunning for NVIDIA in the data center market with a combination of new Xeon Phi CPUs and Altera's programmable chips, which aim to eliminate the need for discrete GPUs in machine learning applications. It's also countering NVIDIA in the automotive market with its Atom automotive processors, Mobileye crash avoidance systems, and Movidius computer vision chips.

Qualcomm (NASDAQ:QCOM) will also become a dangerous rival in the automotive space once its acquisition of NXP Semiconductors (NASDAQ:NXPI) closes, since it will gain NXP's BlueBox autonomous driving platform and become the biggest automotive chipmaker in the world. There are also persistent (but denied) rumors that NVIDIA customer Tesla Motors could work with AMD's partner GlobalFoundries on a next-gen chip for autonomous vehicles.

As for the cryptocurrency market, the overall market remains small compared to the gaming, data center, and automotive markets, and its long-term outlook remains cloudy.

So is there ever a right time to buy NVIDIA?

NVIDIA has repeatedly proven the bears wrong in the past, as its "best-in-breed" chips kept rivals like AMD and Intel at bay. It might continue to do so next year, but the stakes are getting higher with AMD and Intel investing a lot more money into countering NVIDIA.

I believe that NVIDIA's dominant market share in discrete GPUs and its first mover's advantage in data center GPUs and cars should prevent Intel and AMD from gaining much ground.

But I also think that NVIDIA's valuations are too lofty for investors to start a full position at these prices. Therefore, it's smarter to buy a partial position and gradually average into this volatile stock over the next few quarters.

This article represents the opinion of the writer, who may disagree with the “official” recommendation position of a Motley Fool premium advisory service. We’re motley! Questioning an investing thesis -- even one of our own -- helps us all think critically about investing and make decisions that help us become smarter, happier, and richer.