Nvidia (NVDA 2.49%) scored some major victories over the past year with robust sales of gaming GPUs and datacenter GPUs for machine learning. But the chipmaker's biggest win this year so far has been its sales of Tegra CPUs for connected cars. Let's see how Nvidia expanded into this market, and why it could evolve into a new pillar of growth for the company over the next decade.

Image source: Nvidia.

Why Nvidia's chips power cars

Nvidia originally introduced its ARM-based Tegra CPUs to challenge Qualcomm (QCOM 2.01%) in mobile chips. But after Qualcomm conquered the smartphone and tablet markets, Nvidia pivoted the Tegra toward new markets like Android consoles, set-top boxes, and infotainment systems for high-end cars.

Nvidia's Tegra-powered VCM (visual computing module) platform appealed to automakers like BMW, Audi, Honda and Tesla because it delivered better graphics, audio, and 3D navigation for infotainment systems than older ones running on weaker processors. As a result, Nvidia's automotive revenue rose 47% annually to $113 million last quarter, accounting for 71% of its Tegra sales and 9% of Nvidia's top line.

Nvidia is now building upon that foundation with Drive PX, a Tegra X1-powered supercomputer which powers ADAS (advanced driver assistance systems) for "autopilot" features like automatic braking and lane changes. Its successor, the PX2, will reportedly be ten times faster (capable of 24 trillion operations per second), and power fully autonomous vehicles. The next-gen platform has already scored design wins from Volvo, Ford, Daimler, and Audi, indicating that it could become a major platform for driverless vehicles.

The Drive PX2 board. Image source: Nvidia.

Nvidia has competitive advantages

Nvidia's only meaningful rivals in this space are Mobileye (MBLY), an ADAS vendor which sells EyeQ computer vision processors, and NXP Semiconductors (NXPI 2.19%), which recently released the BlueBox turnkey platform to compete against Drive PX.

Nvidia has two major advantages over Mobileye. First, Tesla recently announced that it wouldn't buy Mobileye's EyeQ chips -- which power its Autopilot feature -- beyond its current-gen EyeQ3 chips. That relationship likely ended due to a fatal crash involving a Model S running on the Autopilot feature, and indicates that Tesla could potentially start using Nvidia's Drive PX platform instead. Second, Nvidia finished last quarter with $4.8 billion in cash and marketable securities, while Mobileye had just $543 million. This means that Nvidia can invest more heavily than Mobileye in improving its chips, then possibly sell them at more competitive prices by reducing costs with higher levels of production.

NXP is a tougher opponent. It became the biggest automotive chipmaker in the world after acquiring Freescale last year. Therefore, NXP can probably use scaling and bundling strategies to steal market share away from Nvidia. However, Nvidia's first mover's advantage and established reputation with high-end automakers could keep NXP at bay.

That first mover's advantage might widen Nvidia's moat against other challengers like Qualcomm, which followed Nvidia into the automotive business with its Snapdragon 820 Automotive processor earlier this year, and Intel (INTC 0.85%), which provides infotainment chips for select Hyundai, BMW, Kia, and Infiniti vehicles. Intel also recently struck a deal with BMW and Mobileye to develop new autonomous technologies.

How much bigger could Nvidia's automotive business grow?

PWC estimates that the value of the connected car market will grow from $34.5 billion to $126 billion between 2015 and 2020. PWC divides that market into six categories -- driver assistance, safety, entertainment, well-being, vehicle management, and mobility management. Here are the projected compound annual growth rates for each of those markets:


2015-2020 CAGR

Value in 2020

(billions of USD)

Driver assistance












Vehicle Management



Mobility Management



Source: PWC, converted from euros.

Nvidia can tap into all of these markets to varying degrees with its VCM and Drive platforms. So if Nvidia's Tegra sales grow at a rate comparable to PWC's forecast for the whole connected market, its revenue might rise from $559 million last year to over $2 billion by 2020. That growth would diversify Nvidia's entire business away from GPUs and offset any seasonal fluctuations in GPU sales.

The key takeaway

The transformation of Tegra from a failed Qualcomm competitor into an automotive powerhouse was a game-changing strategy for Nvidia. That move, along with its GPU advances into supercomputers and machine learning, could make Nvidia a much more versatile chipmaker than its traditional rivals AMD and Intel.