THE CONTENDERS

ICE v BEV

The Car replaced the Horse, the Auto Industry evolved on the back of the genius mechanical engineering that was the Internal Combustion Engine (ICE). This new giant was substantially enabled and supported by the ever-growing Oil Industry.

Cars changed lives, and lifestyles. Dominant Auto Industry Brands enjoyed over 100 years of growth and prosperity building “Up and Down” supply chain businesses globally. Around 110 years of progress and complacency these established business giants are now substantially threatened by the introduction of the Battery Electric Vehicle (BEV or EV) which calls upon software and data networking to produce a totally compelling alternatives to cars that are dependent upon the ICE - Internal Combustion Engine

This is the dawn of a revolution in personal transport

Autonomous Vehicles - AVs

These new Software Defined and Fully Connected BEVs provided the platform for further evolution towards Autonomous Vehicles (AVs). The prospect of an autonomous driverless car has morphed from science fiction to fact in the last few years. Every auto Industry participant understands they need to deliver on this compelling new AV Tech... But can they? In 2012 Tesla commenced the disruption with a compelling BEV and continue to take share through innovation. Tesla technology today delivers Full Self Driving (FSD) as an OTA downloadable software. The InCar compute and sensor system has been present in most models since 2019.

Most cars today deliver varying degrees of Driver Assist systems - referred to as ADAS. Advanced Driver Assistance Systems are often confused with Autonomy or claimed to be Autonomous.

ADAS does not equal Autonomy

Building a Safe Autonomous Vehicle is a Very Difficult Challenge

The graveyard of autonomous vehicle aspirations includes most legacy automakers.

Ex GM CEO defines the AV market today - Video

Tesla & Google (Via Waymo) Remain - but crucially only one of them makes cars. Waymo Operates in the Google “Other Bets” division with an analogue Implementation. They have limited stock (2k) of the old Jaguar I-PACE, which they convert at a rate of 60 per week Tesla produces ~35,000 cars per week. All of them are Robotaxi ready - Ex Factory

The existing ICE platforms on which current cars are built are not technically capable, they are neither ‘Software Defined’ or ‘Fully Connected’.

Many Legacy Automakers concluded the liability exposure of Autonomy was too much of a threat.

The Direct to Consumers model for Rides was not attractive, dealers added no value, nor could they be involved.

GM acquired ‘Cruise Automation’ & shut it down following a severe pedestrian injury.

Uber Autonomy Shut down AV efforts following a fatality.

APPLE Invested > $10Bn... only to shutter the project- It’s a very difficult problem.From 2014 to 2025: 8 from 10 declared a shut down of AV efforts.

The AV Industry Has Evolved Into 2 Different Technology Approaches

AV. Digital AI

Probabilistic

$37,000

V

AV. Analogue

Deterministic

> $150,000

Digital AI: Tesla

End to End AI Neural Net

Deep Computer Training

Massive Compute Farms

Factory Line production

> 30,000 per week

Analogue: Everyone Else

Requires Location Mapping

Mapping Maintenance

Heuristic C++ Programming

If This ...Then That

Modify ‘Other’ Vehicles

Converting ~ 60 Units per week

Built In: Digital

(Factory Line)

TESLA ~ $35,000 cost to produce

Bolt On: Analogue

(Post-Production)

V

*

WAYMO ~ $150,000 cost to produce

Vertical Full Stack Integration

Hardware, Software & Sensors

All Designed & built In House

Producing ~1.5m units Annually

Multiple Competing Suppliers

Hardware +Software + Sensors

Bolted Onto or Into A N Other Car

Requires Sensor Maintenance

Converting 60 per week

*Whilst not always technical correct, 'Analogue' v 'Digital' provides a useful nomenclature

AI Neural Network Data Training

In Car Digital Inference Processing via an AI Neural Net provides 50 Trillion operations per second (FLOPS) - In Car
Proprietary Training Data of billions of miles from an 8m Fleet, deployed at an AI training farm of 120,000 Cohesive GPU* Stacks.
This ‘Firmware Factory’ produces FSD-S & FSD-U for OTA downloads to each participating/qualifying car.
No Pre-Mapping of operating area and no maintenance of mapping required.
Production capacity: ~2m Robotaxi compliant cars Per Annum.
All new cars ex factory today are fully Robotaxi equipped
Immense scalability for which there is NO Industry Equal

GPU* =Graphical Processing Unit (Advanced CPU)

Image of 8 GPU Stacks Versus 120,000 deployed

$10Billion invested for FSD-S & FSD-U Training

Just, the cooling for the Firmware Factory!

But How?
AV Digital AI – Building the Competence for FSD Unsupervised

The Partnership Scramble

Legacy Auto and existing Ride Hail Operators have been convinced by supply chains that important missing AV components and skills can be delivered through Partnerships. Existing Ride Hail companies (Uber) believe they can provide driverless services by sourcing cars and components and software from these partnerships by funding their endeavours. It’s a very messy ‘mish mash’ of incohesive multiple 3rd parties.

New Partnerships are aspiring to Compete in the lucrative AV Robotaxi Market but none of them manufacture cars. It is an entertaining bun fight of unrealistic dreams. Waymo and others have proven that 'Analogue' technology can work safely but the unit cost economics & the financial - logistical burden of delivery is a challenge. It makes good PR & gets good press coverage but the profitability, and therefore the viability of this business case is yet to be proven.