Name: Deconstructing Nvidia’s Vera Rubin — The Successor To Blackwell That’s 10x More Efficient
Uploaded: 2026-03-02T22:07:03.000Z
Duration: 14 min
Description: Thousands of YouTube videos with English-Chinese subtitles! Now you can learn to understand native speakers, expand your vocabulary, and improve your pronunciation...

This is what Nvidia's Grace Blackwell looks like on the inside.

The 72 GPU system that's seen soaring sales from the likes of Microsoft, Google, Amazon and Meta.

Adverbs of degree

But what we're really at Nvidia for is an exclusive first look at this.

Nvidia's next RackScale system for AI data centers has the AI world buzzing because Nvidia says it'll help solve the biggest bottleneck threatening AI buildouts today.

It's actually going to be about 10 times more performant in terms of performance per watt compared to Blackwell.

We're taking you on a deep dive inside Vera Rubin to find out.

Because this AI powerhouse takes a lot more companies than Nvidia and a lot more parts than these GPUs.

It also has the compute trays, the chassis, the side rails, the bus bars and 1.3 million components.

Over 80 different suppliers and across more than 20 countries.

VR Rubin's in volume production today, with plans to ship later this year.

But from a memory shortage to a supply chain ridden with tariffs, what are the risks Nvidia faces to getting these two ton systems out the door?

There is that risk that with all this AI capex, you basically have pulled forward or have a little bit too much than you need.

But I don't see that as being a risk this year.

And I think that's why Nvidia is so confident that their customers are saying we need more and if we had it, we could deploy it right now.

From compute to connectivity to power, we take a look at the astounding ecosystem behind this single system and asked Nvidia how it built this complex web and whether competitors like AMD could ever catch up.

When Nvidia released its current Blackwell rack design in 2024, demand was insane.

Nvidia stock is up more than 100% since before it was announced.

Because it was the first full rack scale design that took this approach of disaggregating compute, the scale of networking as well as all the memory infrastructure.

But all of it was done to make it behave as a single gpu.

The first generation of compute knew you'd buy a server, you'd buy a switch, you know you would, you would maybe buy a rack to cool if you got to a point where it needed that.

Got all these systems pulled together into a single rack, built for absolute greatest efficiency and greatest performance.

And that's just not how Servers were historically built.

Grace Blackwell has 72 GPUs, nearly 800 other chips, 1.2 million total components made at some 350 factories.

TSMC is developing the silicon and the core chips.

We're working with Foxconn on a lot of the other rack assembly components.

Delta Electronics for a lot of liquid cooling elements.

The nozzles that connect on the end of the hoses are made by a number of different providers.

Whether that's Amphenol and getting the connectors and the copper, whether that's Vertiv and getting the cooling distribution systems.

With so many components, not every rack is touched by the exact same global suppliers, from China to Israel to Mexico, the us, Vietnam, Thailand.

So Nvidia created a standard reference design.

So the whole idea by creating a standard is it opens it up to the entire ecosystem.

There's PowerShells by Megmeet Lighton or Flex Power components by Infineon analog devices or STMicroelectronics.

Chassis by Foxconn or Interplex Bizlink for bus bars, Pinda for rack manifolds, Coldplates by Auras ABC Boyd Cooler Master JPC to Rico Deal for power Whips.

This supply chain has really become this core group of companies that are building this all together.

And of course, they're all leaning in on the success of Nvidia.

We're still, you know, producing thousands of racks per week.

And so what I'll say is they're spoken for.

So now Nvidia is making the next big leap to Vera Rubin.

With full production announced at CES in January.

This is a Rubin Pod 1152 GPUs in 16 racks, with about 100,000 more components than Grace Blackwell.

V Vera Rubin uses about double the energy, but can deliver way more compute.

The number of tokens it's generating is exponentially higher than last generation.

And so they've become very efficient for these, you know, these heavy workloads.

To get there, Nvidia redesigned all six of the core chips.

It delivers about 2x the performance per watt compared to our previous Gen Grace CPU.

This can actually deliver about 50 petaflops of AI performance, so about 2.5x performance.

It has 18 compute trays, each one starring two Vera Rubin superchips.

Each has one Vera CPU, two Rubin GPUs and much more.

So this particular superchip has 17,000 components.

And as you look at all these elements that have to connect to each other, a lot of these are just to help make sure that the power is is being modulated correctly.

So we actually have memory units that can be applied and removed on the new Vera Rubin processor.

Compared to the Grace Blackwell they were all sort of soldered in.

But the real memory star lines the top and bottom of the Rubin GPU itself.

Eight stacks of the latest high bandwidth memory HBM4 by the likes of SK, Hynix and Samsung.

Is there any risk because it is in such short supply right now?

The one thing that we've been very laser focused on is making sure our entire supply chain is aware and aware of not just sort of the design itself, but

we're actually giving them very detailed forecasts.

We're aligning to make sure that everything we're shipping will be met by our supply chain.

Another risk factor, those 1.3 million components running workloads 24.

Some customers reported overheating issues in the early days of Blackwell deployments.

Every component has to fit together perfectly.

And I think when we were doing some of those initial deployments there were things like you didn't get proper seating with the liquid cooling valve.

It was just actual implementation, user error.

And for the most part all the systems are fully deployed today, running hundreds of thousands of scales and without an issue.

Vera Rubin compute trays have no hoses, cables or fans.

It's Nvidia's first system that's 100% liquid cooled.

Will this mean that data centers have to have a really robust liquid cooling loop ready and in place before they can move from Blackwell to Vera Rubin?

Essentially we are requiring as we build out the AI factories of the future that the vast majority will be leveraging this liquid cool based architecture.

The superchip is covered with cold plates piping in water that keeps the components cool.

This is the cool cold plates on the Vera CPU.

You also have the Rubin GPUs that are also underneath there.

Data centers currently consume a lot of water.

One of the counterintuitive savings is actually you use less water using liquid cooled systems.

The main reason is because you don't use the evaporative cooling technologies as much.

And that allows you to consume much less water through a closed Loop system.

A lot of the needs for cooling come because of just how much power is running through the system.

It's about twice the amount of overall power, roughly about 220 kilowatts per rack.

And so we had to redesign obviously the power delivery system.

But most of all, efficiency is tied to how quickly each GPU can access the memory and processing on every other GPU simultaneously, which Nvidia solves with its own networking product, NVLink.

This is the NVLink switch chip responsible for connecting all of these GPUs and CPUs together to have them behave as one.

So this actually doubled the line rate from again 1.8 terabytes per second to 3.6 terabytes per second.