NVIDIA's (NVDA -3.87%) stock rallied nearly 1,900% over the past five years as shares of Intel (INTC -1.60%) advanced about 140%. NVIDIA's massive gains boosted the chipmaker's market cap to over $150 billion.
That's still lower than Intel's market cap of $240 billion, but NVIDIA's growth trajectory suggests that it could eventually overtake Intel. Here are the top four reasons NVIDIA might pull that off within the next few years.
NVIDIA CEO Jensen Huang. Image source: NVIDIA.
1. Stronger revenue and earnings growth
Wall Street expects NVIDIA's revenue to rise 27% to $12.4 billion this year, and for its earnings to grow 35%. That growth should be fueled by strong demand for its GPUs for gaming, professional visualization, and data centers, along with rising sales of its Tegra CPUs for connected cars and other devices.
Analysts expect Intel's revenue to rise just 4% to $65 billion this year, and for its earnings to improve 3%. Higher sales of its data center chips, along with the growth of its Internet of Things (IoT), non-volatile memory, and programmable solution chips, are expected to offset softer demand for its PC chips.
2. More balanced growth
NVIDIA offers more balanced growth than Intel. Last quarter, its gaming revenues rose 29% annually to $1.74 billion, as its data center revenue -- lifted by the use of GPUs for machine learning purposes -- more than doubled to $606 million. Its professional visualization revenues also grew 13% to $254 million.
NVIDIA's two slower-growth units are its automotive and OEM/IP businesses, but neither unit is shrinking. Its automotive revenues rose 3% annually to $132 million last quarter, while its OEM and IP revenues advanced 2% to $180 million. That across-the-board growth boosted its total revenue by 34% to $2.9 billion for the quarter.
Intel's soft spot is its "PC-centric" Client Computing Group, which saw its revenue dip 2% annually to $9 billion last quarter on tougher competition and slower upgrades.
Meanwhile, its combined "data-centric" (IoT, memory, and programmable chips) revenues rose 21% to $7.9 billion, but much of that growth was offset by its slumping PC revenues. As a result, its total revenue only rose 4% to $17.1 billion for the quarter.
3. Stronger footholds in the right markets
Intel dominates the data center market with its Xeon x86 CPUs, which power about 99% of all servers worldwide. However, data center CPUs aren't as frequently upgraded as PC CPUs.
Image source: NVIDIA.
Meanwhile, NVIDIA established a firm foothold in the data center market with its Tesla GPUs, which are generally more important than CPUs for processing machine learning and AI tasks. It also aims to disrupt the data center market with products like the $149,000 DGX-1 "supercomputer", which offers the power of 400 servers in a single box. The DGX-1 is powered by eight Tesla V100 GPUs and two 20-core Xeon CPUs.
Worldwide sales of desktop PCs remains weak, but demand from PC gamers remains consistently high, due to strong sales of graphically demanding games. GPUs are generally more important for high-end games than CPUs. That's why NVIDIA, the top maker of discrete GPUs in the world, didn't struggle as Intel repeatedly blamed its sluggish CPU sales on soft PC sales.
In the automotive market, NVIDIA offers its Drive PX series of onboard supercomputers, which convert regular vehicles into driverless ones. Intel is still developing a rival platform via a partnership between its subsidiary Mobileye and automakers BMW and Fiat Chrysler.
4. A longer technological runway
Intel is running out of room to make better chips. In 2016, it killed off its two-year "tick-tock" cycle, the basis of Moore's Law, and replaced it with a three-phase "Process/Architecture/Optimization" cycle over a period of almost three years. This means that it's taking longer for Intel to make faster chips -- which means that its core PC and data center customers will likely postpone their upgrades.
NVIDIA, on the other hand, has a much clearer roadmap. It launched its next-gen Volta-based GPUs for workstations last year, and will reportedly follow that up with two brand new GPU architectures -- Turing for gaming and Ampere for enterprise (or vice versa, since the names aren't finalized) -- later this year.
These moves are all happening as NVIDIA's main GPU rival, AMD, still struggles to match the performance of NVIDIA's current-gen Pascal GPUs at comparable prices.
Intel sees the writing on the wall
NVIDIA is slowly creeping up on Intel across multiple markets. Although Intel doesn't view NVIDIA as a direct competitor, its CPUs are becoming less relevant than NVIDIA's GPUs in its core PC and data center markets.
That's why Intel started developing its own discrete GPUs late last year. However, that eleventh-hour effort simply indicates that Intel is scared of NVIDIA -- and likely knows that NVIDIA could eclipse its market cap within a few years.
