Audi and Aptiv invest $285 million in ‘orchestral director’ for self-driving car data

U.S. automotive parts supplier Aptiv joined Samsung Electronics, German chipmaker Infineon, and Volkswagen Group premium brand Audi as a major shareholder in Austria’s TTTech Auto. 

Unlike VW associate Argo AI, the Vienna-based firm doesn’t design algorithms to navigate cars through their environment. Instead it ensures mission-critical data circulates throughout the vehicle reliably and swiftly to minimize safety risks—an increasingly complex task as computing requirements of assisted and automated driving features soar.

“Think of us as a kind of orchestral conductor managing the flow of information,” chief executive Georg Kopetz told Fortune. “A sensor signal detecting a potential crash has to be assigned the highest priority in the car’s IT network, so a command can be sent to activate the brakes in just milliseconds.”

Much like ensuring an airplane’s electronics function under the most demanding fail-safe requirements, it’s the kind of inner plumbing that only gains notice when something goes wrong. Without being able to guarantee this level of safety, automakers will be loath to market automated driving features.

Courtesy of TTTech Auto

Its main product is MotionWise, a software program akin to an operating system. It is designed to be microprocessor agnostic while simultaneously meeting the highest industry IT fail-safe standard known as ASIL-D.

That means its software can be continuously developed to include new features and remotely updated, independent of the hardware on which it runs. Once in operation, information packages can then be shifted back and forth seamlessly between processing cores depending on their safety priority. Data from the airbag control unit would be more important than that of an infotainment system like a center console display.

This allows automakers to pick and choose their chip suppliers—whether Intel, Qualcomm, Nvidia, or Samsung—depending on their computational needs, without problems arising in software integration.  

“Decoupling the software from the computing core makes it quicker to deploy to the fleet and easier to scale,” Kopetz said. “This is called abstraction, and a similar approach can be found, for example, with VMware in the server industry.” 

‘Democratize’ automated driving

Aptiv, Audi, and TTTech Auto have collaborated before, with the trio teaming up to develop the Audi A8’s central brain, initially designed to power unsupervised “eyes-off” driving in late 2017. 

The feature was so advanced at the time that it would take regulators until June 2020, nearly three years later, before there was even a method to approve the feature for sale, so it never came to deployment.

The investment, split with 20% from Audi and 80% from new shareholder Aptiv, values the company at $1 billion, according to J.P. Morgan, TTTech Auto’s financial adviser on the deal.

Aptiv CEO Kevin Clark said the investment would accelerate development, testing, and validation of its automotive products, enabling the company to help “democratize advanced safety systems faster and at a lower cost.” 

The key word is democratize: Currently any vehicle offering even unsupervised automated driving is out of the price range of all but the most affluent car buyers.

Once you get to robotaxis like the Cruise Tostada currently roaming the streets of San Francisco, the cost becomes prohibitive owing to the amount of sensors, compute power, and redundancies built into the vehicle.