NVIDIA Confirms Vera Rubin Launch In Q3 With Volume Ramp by Q4, As Blackwell Continues To See Massive Demand

NVIDIA confirms its next-gen Vera Rubin AI platform timeline, but existing GPUs continue to see major traction from AI firms, leading to price hikes.

NVIDIA Vera Rubin Is All Prepped To Power Agentic AI Factories In Q3, Volume Ramp In Q4 2026 & Bigger Numbers In Early 2027

Today, NVIDIA announced its Q1 FY2027 earnings with a record revenue of $81.6 billion, up 85% from the last year. The revenue was driven primarily by the Data Center segment, which held the mammoth share of $75.24 billion, while Edge Computing amounted to $6.3 billion in revenue.

Revenue by Market Platform

(in millions of dollars)

Category	Q1 FY27	Q4 FY26	Q1 FY26	Q/Q	Y/Y
Data Center	75,246	62,314	39,112	21%	92%
Hyperscale	37,869	33,814	17,599	12%	115%
AI Clouds, Industrial, & Enterprise	37,377	28,500	21,513	31%	74%
Edge Computing	6,369	5,813	4,950	10%	29%
Total	81,615	68,127	44,062	20%	85%

During the earnings call, NVIDIA highlighted sustained demand for its Blackwell GB300 & NVL72 systems, which continue to see huge momentum in frontier models and hyperscalers. Currently, 100s of 1000s of Blackwell GPUs are deployed across the world, with the total number of data centers exceeding 10 Mega-Watts, now sitting at 80 sites. NVIDIA's CFO states that Blackwell has been their fastest product ramp in the company's history, and its demand is propelled by its "lowest token generation cost" at inference.

The demand for GB300 and BL72 was particularly strong, with frontier model builders and hyperscalers each having cumulatively deployed hundreds and thousands of Blackwell GPUs, marking the fastest product ramp in our company's history. Brace Blackwell is the fastest training system as well as the lowest token generation cost at inference.

The number of partner data centres exceeding 10 megawatts has nearly doubled in just one year, now surpassing 80 sites.

Colette Kress - NVIDIA CFO

At the same time, NVIDIA is seeing a price surge across older GPUs too. Jensen recently said that this was the "Fine Wine" effect. As GPU shortages continue to grip the AI segment, prices across older GPUs have seen an uptick.

Older GPUs such as the Hopper H100 are up in price 20%, while the Ampere-based A100 is up by nearly 15%, versus the previous year. Cloud providers are further opening up access to AI firms through their older hardware stacks as AI demand rages on.

The price of renting an H100 has risen 20% year-to-date, while A100 cloud pricing is up nearly 15%.

Colette Kress - NVIDIA CFO

The Vera Rubin Storm Is Coming

But Blackwells, Hoppers, and Amperes are past news; all eyes are now set on NVIDIA's next-generation Vera Rubin platform, which is part of its Extreme Co-Design ecosystem, which spans multiple chips and multiple software solutions, as the one-stop shop for all your Agentic AI needs.

Vera Rubin is now in production, and NVIDIA plans to ship the first shipments of Vera Rubin by Q3 2026. This will be followed by a volume ramp in Q4 2026 and increased shipments through the first half of 2027. NVIDIA's CEO states that the pace of their annual product cadence is unmatched, and following Rubin, we will see Rubin Ultra by 2027 and Feynman by 2028.

Do note that NVIDIA is also mass-producing Vera CPUs, which are on track for deployment in racks. The first shipments of Vera CPUs have already rolled out to early customers such as OpenAI, SpaceXAi, Oracle, and Anthropic.

Our annual product cadence, a pace that is unmatched, remains a key pillar supporting our market position. We are on track to commence production shipments of Vera Rubin in the second half of this year, starting in Q3.

[...]Yeah, well, we have indicated for a while that we will be launching Vera Rubin in the second half. We will start in Q3. That will be our initial pieces together. And then once we get to Q4, we're probably going to start to see our ramping continue. It's hard to say at this point what will be a faster ramp. But again, we have demand already planned.

Jensen Huang - NVIDIA CEO

Talking about the growth segment, inference is where all the action is for NVIDIA. CEO Jensen Huang said that they are growing their inference share rapidly, and that Vera Rubin is already off to a "Tremendous Start," and it will be far more successful than Grace Blackwell. We have seen the initial performance figures, and we have also looked at the demand for Rubin across various AI firms, so that much should be true.

Well, we are growing share in inference, and we're growing share in inference very, very quickly... And so we're gaining share tremendously fast in inference. Vera Rubin is going to be even more successful than Grace Blackwell at this point.

Every single, I can't think of one, every single frontier model company will jump on Vera Rubin from the get-go. And that wasn't true before on Blackwell. And so Vera Rubin is off to a tremendous start, and it'll surely be more successful than even Grace Blackwell.

Jensen Huang - NVIDIA CEO

NVIDIA is firing on all cylinders. While its current Blackwell platform delivers record-breaking revenue and continues to see explosive demand — even driving up prices for older Hopper and Ampere GPUs — the company is already gearing up for the future.

The next-generation Vera Rubin platform is on track for initial shipments in Q3 2026, with volume ramping in Q4 and strong growth into 2027. Designed to power the coming wave of agentic AI factories, Vera Rubin is poised to build on Blackwell’s success and further solidify NVIDIA’s dominance in the AI infrastructure race. The AI boom shows no signs of slowing down, and NVIDIA is ready for what’s next.

Follow Wccftech on Google to get more of our news coverage in your feeds.