Microblog: Dylan Jovine – NVIDIA Partner for Autonomous Driving Using THOR Chip

Anybody have ideas about who this could be? He talks about a tiny, unknown CPU maker that is partnering with NVIDIA on their THOR Chip to make autonomous driving level 4 (completely unmanned).

Some text from Dylan Jovine’s teaser…

“What makes THOR so powerful is that it’s designed for “level 4 autonomy.”

That means it can fully power autonomous systems with no human oversight.

Of course, that depends on where it’s legal, but that is what Nvidia’s new superchip is capable of.

Now, let’s compare THOR to what’s on the market…

Right now, cars need lots of separate chips for advanced driver assistance, infotainment, instrument clusters – the list goes on and on.

If you’ve been in a Tesla or Mercedes or Porsche or any of these fancy cars with driver assistance, or an autonomous vehicle…

There’s a lot of stuff going on, okay?

For one, you have all the cameras – visual and infrared cameras.

Then you have LiDAR laser depth mapping…

Radar for velocity and distance…

Ultrasonics for close range…

This whole stack to process surroundings – people walking on the street, lights changing, other cars coming.

It’s a lot to process.

And all these things have been processed by individual chips.

But THOR is actually going to do all of it.

Now, here’s the interesting part –

The part NOBODY is talking about.

Nvidia did something very clever here…

They took a CPU from a little-known company we’ll be talking about on our 1:00 ET webinar tomorrow [CLICK HERE TO RSVP]…

So they took this CPU – central processing unit – from this “secret supplier” – and it actually manages operating system tasks, sensor synchronization and decision logic…

And combined it with Nvidia’s GPU – graphic processing unit – which processes for perception and object detection.

So it’s like its own little computer.

So you have the CPU chip processing all the decisions, logic, synchronization…

While Nvidia’s chip is processing the environment – the mapping and all that stuff.

Now, listen – this is why we have a situation where this whole autonomous vehicle game is about to take off…

There have been two major bottlenecks the past two years that have stopped this from exploding so far…

One has been the chips problem – the processing problem: too many chips, not fast enough.

Bottleneck number two has been data centers.

Remember, there’s a lot of data running through the car itself, but also pulling for object recognition, etc.

Well, these two issues have been resolved.

That’s why Nvidia CEO Jensen Huang calls this “a $50 trillion market.”

And there’s no doubt about it.

Just to give you some context – there’s 2.3 billion vehicles on the road around the world.

The United States has about 350 to 360 million.

In 10 years, 30% of those vehicles will have fully autonomous capability.

That’s 690 million fully autonomous cars.

And it all starts with Nvidia’s new chip – THOR.

Now, the really, really big story here that no one is talking about is the secret supplier behind these “Nvidia” chips.

This is a discussion topic or guest posting submitted
by a Stock Gumshoe reader. The content has not been edited
or reviewed by Stock Gumshoe, and any opinions expressed
are those of the author alone.