BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Intel Acquiring Mobileye For More Than $15B In Major Move Into Automated Driving

Following
This article is more than 7 years old.

As the automotive industry races to commercialize automated driving by the turn of the decade, the number of real power players in the space is rapidly consolidating. The latest move is Intel’s acquisition of Israeli machine vision leader Mobileye for $15.3 billion. If completed, the deal may finally put Intel in a strong position in a new market after it has struggled in recent years to break into new territory beyond its deteriorating stronghold in PC chips.

Founded in 1999, Mobileye is the undisputed global market leader for the image processing systems that are increasingly becoming standard equipment on new vehicles and will be essential to making self-driving vehicles work. Mobileye doesn’t actually make the cameras used for lane-keeping and pedestrian detection systems, but it designs the software algorithms that recognize the objects that are visible in each frame captured by the camera as well as the chips that power the software. Most of the world’s major automakers use Mobileye technology.

Intel has been the dominant manufacturer of general purpose microprocessors since introducing the technology in the 1970s. However, in the past decade as the market has shifted from traditional laptops and desktops to mobile devices, Intel has been rapidly losing market share to chips based on more power-efficient designs from ARM Holdings. Intel’s own mobile chip offerings have been a failure in the marketplace and have been discontinued.

One of Intel’s continuing strengths has been in the server space, where it’s high-end chips have the ability to run multiple tasks simultaneously in secure virtual machines. In recent years, Intel has been adapting this capability to automotive-grade processors and promoting new chips to the industry as a means to consolidate the number of electronic control units (ECUs) in new vehicles. Today’s most advanced vehicles may contain as many as 70 to 100 discrete computers or ECUs, all of which require a lot of interconnected wiring and provide potential attack surfaces for hackers.

Navigant Research

Navigant Research projects that by 2025 there will be more than 4.5 million vehicles sold annually with at least Level 4 automation that lets the vehicle drive itself completely in many conditions and at least 70 million with Level 2 semi-automated capabilities similar to Tesla AutoPilot.

Automotive supplier Delphi has been promoting its Multi-Domain Controller (MDC) concept for several years. In November 2016, Delphi announced a partnership with Intel and Mobileye to commercialize the MDC using chips and software from the two companies. Mobileye’s next-generation EyeQ4 and EyeQ5 chips would be combined with one of Intel’s new automotive processors in the Delphi automated driving MDC.

Mobileye’s chip design is specifically optimized for machine learning applications and will be used to process and fuse all of the signals coming into automated vehicles from cameras, LiDAR, Radar and ultrasonic sensors as well as the more traditional wheel speed sensors and accelerometers used for current stability control. The fusion of all of these inputs into a coherent view of the world around an automated vehicle will provide situational awareness for the control system. That control that determines when to accelerate, brake and steer and manage all of the actuators in the vehicle will run on the Intel chip.

This is similar architecture to the one designed by competing chip-designer Nvidia for its Drive PX2 platform. Nvidia uses one of its high-end graphics processing units for the machine learning and sensor fusion with a more general purpose chip for path planning and control. At the 2017 CES in Las Vegas, competing suppliers ZF and Bosch both announced plans to commercialize the Nvidia technology.

Delphi is developing its automated driving stack using Mobileye and Intel technology independently with plans to offer a turnkey solution to automakers by 2019 using software from its Ottomatika division and a suite of sensors. Intel and Mobileye have a separate deal with BMW to supply chips for the German automaker’s automated driving solutions.

Mobileye also faces increasing competition in the automotive image recognition marketplace. In addition to the capabilities provided by Nvidia's platform, other big automotive suppliers, including Hitachi, Autoliv and Dura automotive, are grabbing market share. In some cases, automakers themselves are developing their own technology or acquiring it. In mid-2016, Tesla stopped using Mobileye's technology following a fatal accident of a driver using Autopilot and the company's second-generation system was developed in-house. In August 2016, Ford acquired another Israeli startup called SAIPs to do the same thing, and machine vision is believed to be one of the key factors that led GM to acquire Cruise Automation.

As part of the acquisition, Intel will likely take over production of Mobileye chips from French-Italian producer STMicroelectronics. While unit volumes will likely pale in comparison to the mobile device market for many years to come, the potential revenue growth in the automated driving space is much higher. By the end of 2016, fewer than 20 million vehicles in total were equipped with Mobileye technology.

While the market is going to be growing rapidly and there will be strong competition from Nividia, Qualcomm and others, it’s still very early in the game and the Intel/Mobileye pairing has the potential to grab a significant share, something Intel has failed to do in other recent growth markets.

Follow me on LinkedInCheck out my website