Back to top

Why Are Chipmakers So Interested in Autonomous Cars?

Read MoreHide Full Article

It’s the same reason that automakers are interested in the market. Although for them, it’s also a question of preventing tech companies from encroaching on their turf. There doesn’t seem to be any likelihood of such encroachment however, at least in the near future, as there is too much to making a car that’s just too far from what tech companies do best. So the first fully-autonomous cars that make it to the market will most likely be a collaborative effort between auto makers and tech companies.

The primary selling point for self-driving cars is increased safety due to fewer accidents. There has been some research suggesting that when computers take over the roads, they will remove the chances of accidents due to human error (for sleep, distraction, intoxication, carelessness, brake failure or other mechanical problems, or when it’s the other guy’s fault, or any other reason). And of course this is what is broadly propounded by the companies invested in the phenomenon. Kathy Winter, the VP and general manager of the automated driving division at Intel (INTC - Free Report) thinks that it will also help create logistics solutions for transportation and delivery of products.

But don’t expect a sudden jump from driven to driverless cars. It’s more likely that cars will increasingly incorporate features that automate certain functions over time, making them semi-automated to varying degrees. Regulatory and insurance considerations will also likely require the presence of a driver long after cars become fully-automated.

The Institute of Transportation Studies at University of California, recently conducted a survey of forty experts including policymakers, researchers, government representatives, nonprofit organizations and tech companies, presenting the results in a report titled "3 Revolutions: Sharing, Electrification, and Automation." The chief findings were as follows:

  • 70% of survey participants think fully driverless vehicles will account for more than 20% of vehicles sold by 2040.
  • 88% think commercially offered shared rides will make up more than 5% of all U.S. passenger miles by 2030, and 78% think commercially offered shared rides will account for more than 20% of U.S. passenger miles traveled by 2040.
  • 70% also think that by 2050, the majority of vehicles used commercially for ride and car sharing in the nation will be zero emissions vehicles, including battery, plug-in hybrid, and fuel-cell electric vehicles.

According to IHS Automotive forecasts, several thousand autonomous vehicles will be on U.S. roads in 2020, with the number growing to nearly 4.5 million vehicles by 2035. By then, there will be 21 million autonomous vehicles on the road globally, growing at a 43% CAGR from 2025 to 2035.

No wonder everyone is excited.

What’s In It For Chipmakers?

Both Intel and NVIDIA (NVDA - Free Report) say that the fully automated car will have as much electronic horsepower as a super computer.

Whether a lidar or radar system is used to pick up images of the area surrounding the car, the system is based on semiconductors. These images have to be processed and matched with relevant data to decide a course of action at the blink of an eye. So there’s processing (analog, logic), communications and machine learning involved. Each of these operations will be facilitated with semiconductors. Plus there’s entertainment, displays and syncing with other devices that will also require semiconductors.

Key Players

The Intel Go computer and NVIDIA Drive PX 2 are designed to help cars detect pedestrians, objects, lanes, respond to traffic signals, etc. BMW is testing the Intel solution while Tesla (TSLA - Free Report) is testing the NVIDIA solution. Baidu (BIDU - Free Report) is developing an artificial intelligence platform that combines its cloud platform with NVIDIA self-driving technology to deliver things like HD maps, vehicle control, automated parking, etc for the Chinese market.

NVIDIA has a lead in the machine learning segment and its automotive chips are already in millions of cars from a number of suppliers, which in combination, gives it an edge in the automated car segment.

Intel has been in collaboration with Mobileye for quite a while and has finally announced that it will buy the Israeli supplier of self-driving technology (camera and laser-based sensors) for $15 billion. Mobileye currently sells its technology to 27 automakers across the world, so the combination will be highly advantageous for Intel.   

Qualcomm (QCOM - Free Report) , which beat Intel hands down in the mobile computing business, is also set to rule in self-driving cars (although this time, competition will be tough with both Intel and NVIDIA having gained some mileage). The company recently launched its Drive Data Platform, which is its version of a self-driving system to process the sensor inputs generated by a vehicle. The system is based on the Snapdragon 820Am and makes use of the Snapdragon Neural Processing Engine (SNPE).

But there’s more. Qualcomm’s soon-to-be-completed acquisition of NXP Semiconductors in an all-cash deal worth $47 billion brings on board NXP’s RF SoC for in-vehicle infotainment (IVI) systems. While IVI may not sound that exciting in the context of self-driving cars, the thing to bear in mind in this case is the SoC technology that will shrink the size of high power systems while lowering the cost.

This will be quite a breakthrough for Qualcomm as a lower price will speed up adoption of its systems. Additionally, NXP is the biggest supplier of automotive chips worldwide so it has the necessary relationships to jumpstart its business.

Alphabet (GOOGL - Free Report) has developed self-driving technology of its own but its component suppliers aren’t known. Back in 2015, the company quietly acquired Lumedyne, a maker of micro-electromechanical systems (MEMS) sensors including an optical accelerometer, a vibration energy harvester and a “Time Domain Switched” type of inertial sensor for $85 million. Lumedyne also had a number of patents pending approval at the time.

Over the years, Google has also been hiring a growing number of people with chip design skills. In 2015 itself, Google’s Pixel team advertised an opening for a "multimedia chip architect" who can "lead a chip development effort" and "work with other engineers to take chip to product shipment". Google was particularly looking for expertise in "image processing, video processing, stabilization." The Pixel still uses NVIDIA chips, but Google does have internally developed self-driving technology and the greatest number of self-driving miles on record to train its self-driving system.

Wrapping Up

The self-driving/autonomous car conversation is just getting started. There will no doubt be much innovation in the space over the next few years, both by the biggest tech players and automakers and by smaller startups. In fact, CB Insights estimates that financing in the auto tech startup sector was over $1 billion in 2016 and more money will surely follow. This is one of the best places to be in because the technology is just being built, so it could be good business. Also, early investors can sell out their interest at a nice premium.

For chipmakers it’s a rush to grab a share of the pie.

More Stock News: 8 Companies Verge on Apple-Like Run

Did you miss Apple's 9X stock explosion after they launched their iPhone in 2007? Now 2017 looks to be a pivotal year to get in on another emerging technology expected to rock the market. Demand could soar from almost nothing to $42 billion by 2025. Reports suggest it could save 10 million lives per decade which could in turn save $200 billion in U.S. healthcare costs.

A bonus Zacks Special Report names this breakthrough and the 8 best stocks to exploit it. Like Apple in 2007, these companies are already strong and coiling for potential mega-gains. Click to see them right now >>

More from Zacks Analyst Blog

You May Like