"There are up to 70 small computers in a car, controlling different functions from brakes to acceleration, to the way the engine works, to ... instrument clusters, radio - all these systems need a computer to work," said Luis Gargate, a computational physicist and business development manager at Critical Software, a Portuguese company that develops critical safety systems for automotives, railway and spacecraft.
"When you press the accelerator or brakes on a car, there's software in the loop that will brake your wheel. If it fails, it can be catastrophic because your car might be unable to stop." If you layer on new systems required for self-driving cars - radars, LIDAR, cameras, sensors and memory required to process the data from these sensors - cars are becoming one of the most sophisticated machines to have ever existed.
To put this in context, consider this: Apollo 11, the spaceship that took humans to the moon, had 145,000 lines of computer code. The Android operating system has 12 million. A modern car has about 100 million lines of code.
Everyone from traditional carmakers to technology companies like Apple, Google and Uber, and specialist software makers that build the computing brains of cars, all predict that connectivity and automation is the inevitable next step for automobiles.
Apple has hired several high-profile car experts including Megan McClain, a former Volkswagen engineer with expertise in automated driving, and Vinay Palakkode, a graduate researcher at Carnegie Mellon University, a hub of automated driving research.
Meanwhile, Uber has opened a robotics centre in Pittsburgh that poached a number of researchers from the nearby Carnegie Mellon University and is building mapping technology with the University of Arizona, and Google's self-driving cars had driven over 1.6 million kilometres as of June.
Autonomous driving software has existed for many years now outside the car industry - particularly in space satellites and planes.
Planes can land automatically on certain runways in good weather conditions. The problem with cars is they sit in a far more complex environment, with much more chance of human error.
"Companies like Tesla and Google ... [have] proven the concept of driverless cars, but ... these systems are not failsafe yet. You still need a driver to sit there and react if something goes wrong," Mr Gargate said.
In a sense, cars are much more complex than a typical space satellite. Because of this, cars have begun to borrow technologies originally developed for space exploration.
For instance, Mr Gargate has implemented stress test software originally used on European Space Agency's Sentinel 1 satellite, for software that goes to carmakers like Volkswagen, Audi and Mercedes.
The BMW 7 series uses the head-up displays that fighter pilots used as part of the display systems in its machines. "We have a 40GB hard drive in this car just for the entertainment system," BMW UK's Gavin Ward said.
Will the incumbents still exist as cars become the domain of software engineers, rather than mechanical ones, or be totally supplanted by the likes of nimble Tesla and giants like Apple?
BMW's Mr Ward says bring on the competition. "We see Apple and Google as welcome competition, the motorcar revolution needs new blood, it stimulates ... our engineers."
Future vehicles
Apple
The Apple car will reportedly be fully electric with in-car connectivity like the CarPlay system. It should also be capable of autonomous driving
Google
Google's self-driving cars, which use a LIDAR system to navigate, had driven more than
1.6 million kilometres as of June.
Uber
Its new robotics centre, the Advanced Technologies Centre in Pittsburgh, reportedly poached a number of researchers from the nearby Carnegie Mellon University, which specialises in autonomous robotics research.
