NVIDIA (NASDAQ:NVDA) was supposed to be the dominant force in self-driving cars. The graphics specialist was making a lot of noise about how its graphics processing units (GPUs) are ideal for this application, and even landed a big-name customer in the form of Tesla. But all that hype didn't translate into actual growth.
The chipmaker has lost the initiative in autonomous cars to the likes of Intel. But this shouldn't discourage investors from betting on NVIDIA's self-driving car ambitions. The recent setback is temporary, and there are encouraging signs that it could turn this business around. Let's take a look.
Gearing up for the long haul
NVIDIA had to pivot its automotive business to escape commoditization. Earlier, the company was supplying chip platforms for vehicle infotainment systems, but that market was getting more competitive, thanks to the entry of new players such as Intel. NVIDIA would have been forced to lower prices to remain competitive. This is why NVIDIA decided to exit this market and shifted its efforts toward selling end-to-end self-driving systems built using the company's own hardware and software.
There's no doubt that this pivot has hurt NVIDIA's automotive growth in the short run. The company's automotive revenue in fiscal 2018, which ended in late January, increased 15% annually to $558 million. For comparison, it grew at a much faster pace in fiscal 2017 when revenue jumped to $485 million from just over $300 million in the preceding year.
But that isn't the right way to check NVIDIA's automotive progress. The change in NVIDIA's strategy is setting the company up for growth in the long run despite short-term pains. The company now has a lot more strategic partners in its autonomous vehicle ecosystem as compared with the time when it was focusing on selling just chips.
For instance, at the end of fiscal 2017, NVIDIA had just over 100 partners using the DRIVE PX2 chip. Less than half of them were developing self-driving systems using the company's artificial intelligence platform. However, both these numbers have shot up impressively in recent quarters, and almost all of the company's DRIVE customers are not just buying chips, but also engaging with NVIDIA to develop AI systems for their cars.
As such, NVIDIA has been able to create a solid network of potential customers who will buy its autonomous driving platforms once they hit the market. This strategy seems to be working, as NVIDIA's automotive revenue got a nice shot in the arm last quarter after one of its new products hit the market.
Stepping on the gas, gradually
NVIDIA's automotive revenue increased 13% annually last quarter. This might not sound like much at first, but it is definitely a green flag, as the company's automotive growth curve hasn't been in the best shape of late.
This bump in NVIDIA's automotive revenue last quarter can be attributed to the new DRIVE Xavier platform that went into production during the first quarter of the fiscal year. The Xavier system on a chip (SoC) was originally announced in late 2016, but the company started delivering the same to customers from early 2018. Not surprisingly, it seems to be playing a role in the resurgence of its automotive business given the features it packs.
NVIDIA claims that it has spent $2 billion over four years, employing more than 2,000 engineers to develop Xavier. This new SoC will be fused with NVIDIA's AI-enabled DRIVE software stack to bring advanced capabilities to autonomous cars.
For instance, Xavier will enable AI assistants for drivers and passengers by utilizing sensors deployed both inside and outside the vehicle. It will also power an augmented reality interface by integrating computer graphics, AI, and computer vision.
So NVIDIA now has a better chance of boosting its revenue by selling such end-to-end systems in place of just shipping the chip to a carmaker. But this isn't the only card that NVIDIA has up its sleeve, as its highly capable Pegasus platform is expected to go into production this quarter.
Pegasus is supposed to help automakers roll out vehicles with Level 5 autonomy, the highest level of self-driving that removes the need for steering wheels and pedals. NVIDIA is targeting the robotaxi market with this product, and it has landed over 25 customers looking to develop self-driving taxis using Pegasus.
The big opportunity
NVIDIA is playing the bigger game by building a large ecosystem of partners that could eventually buy their complete self-driving system from the chipmaker. This is a smart strategy as compared with the company's earlier strategy of selling chips to automakers -- NVIDIA can command a higher price for an integrated platform that fuses both hardware and software.
More importantly, the self-driving car space is just getting started. Allied Market Research estimates that sales of autonomous vehicles will increase 10 times from 2019 to 2026. NVIDIA is getting its products out in time to tap into this massive growth and has built a solid base of potential clients, making it a good bet on the autonomous vehicle industry despite recent hiccups.