The Revolution in Autonomous Driving: Tesla, NVIDIA, and the Race to AI Supremacy

 In the world of Artificial Intelligence and autonomous vehicles, two giants stand out - Tesla and NVIDIA. Both are pioneering, but they adopt slightly different strategies, which begs the question, who is leading the charge in the autonomous race?

The State of Autonomous Driving and AI

Today, the race to deliver fully autonomous vehicles is as much about AI and computing power as it is about traditional car manufacturing. No one understands this better than Elon Musk and Tesla. They've crafted a strategy to achieve 100x computing power by October 2024. This goal, while ambitious, is no simple feat.

To grasp the scale of this target, imagine combining the power of 300,000 NVIDIA A200 GPUs or 100 Tesla ExaPods. The costs could well exceed a trillion won, and it's a task even beyond TSMC's 7nm process capabilities.

The Vision of Tesla and NVIDIA

Tesla vs NVIDIA

Elon Musk's recent revelations about the Dojo supercomputer illuminate his company's vision. Tesla, unlike NVIDIA, isn't looking to sell computing hardware but aims to provide services using their Dojo as a platform.

Autonomous driving services are the primary target, with language models and new drug development to follow. In a nutshell, Tesla is saying to other companies, "You provide the body, we'll provide the brain." This sets the stage for a fierce competition with NVIDIA.

NVIDIA: A Formidable Contender

 NVIDIA is already providing a comprehensive autonomous driving platform that includes training models, mapping, data labeling, and edge computing. More than 20 out of 30 EV companies use NVIDIA's solutions. Mercedes Benz, for example, is entirely dependent on NVIDIA for autonomous driving.

Data centers can't function without NVIDIA's products either. And with NVIDIA's growing influence in China, it's clear why they're celebrating their success.

The Power of Data

While NVIDIA's hardware and computing capabilities are formidable, Tesla holds an enviable advantage - an overwhelming amount of data. Tesla's FSD (Full Self-Driving) data has grown almost tenfold in less than a year. Tesla is capturing more data than its competitors by several orders of magnitude.

Edge Computing in Autonomous Vehicles

Edge computing, or on-device computing, is crucial in self-driving cars. Real-time decision-making needs to be local to the car to avoid dangerous latency. Tesla's in-car computers are a marvel of design, combining many functions into a single chip, improving efficiency, and outshining NVIDIA's earlier chips used in Tesla vehicles.

Tesla Chip

With rumors of a new 4.0 chip coming to the Model 3, and Musk's bold claim that a 5.0 chip could be ten times safer than a human driver, Tesla's commitment to vertical integration is clear.