Nvidia unveils Vera Rubin, a next-gen AI system with 10x efficiency gains over Blackwell

Nvidia's new AI system Vera Rubin is 10 times more efficient than its predecessor -- here's a first look

First look at Vera Rubin, Nvidia's next AI system that's 10 times more efficient Nvidia's earnings on Wednesday are expected to show booming sales of the company's current rack-scale system. But all eyes are on its next AI system, Vera Rubin, which is scheduled to roll out later this year. Vera Rubin, which is made up of 1.3 million components, will deliver 10 times more performance per watt than its predecessor, Grace Blackwell, the company claims. That's a significant development when energy consumption is one of the most critical issues facing the artificial intelligence build-out. CNBC got an exclusive first look at Vera Rubin at Nvidia's headquarters in Santa Clara, California. Nvidia says the new AI system is a complex web of parts sourced from around the world. Its core chips include 72 Rubin graphics processing units, or GPUs, and 36 Vera central processing units, or CPUs, primarily made by Taiwan Semiconductor Manufacturing Co. The other parts, from liquid cooling elements to power systems and compute trays, come from more than 80 suppliers in at least 20 countries, including China, Vietnam, Thailand, Mexico, Israel and the U.S. One big challenge the company faces is the soaring costs of memory due to a global shortage from all the AI-driven demand. Dion Harris, Nvidia's AI infrastructure head, said in an interview that the company has been giving suppliers "very detailed forecasts." "We're aligning to make sure that everything we're shipping will be met by our supply chain," he said. "We're in good shape." It's a critical moment for Nvidia, which dominates the market for AI processors but faces intensifying competition from Advanced Micro Devices as well as custom silicon from Broadcom and Google's homegrown tensor processing units. Nvidia has plans to manufacture up to $500 billion of AI infrastructure in the U.S. through 2029, including making Blackwell GPUs at TSMC's new Arizona fabs. Grace Blackwell went into production in 2024, and changed the game on how much compute was possible with a single system. Vera Rubin, which is expected to ship in the second half of 2026, takes the company to another level. Nvidia CEO Jensen Huang announced in January that the system was in full production.

Here's a Look at One of the World's Most Complex AI Systems, the NVIDIA Vera Rubin, Integrating a Million Components

NVIDIA's next-gen Vera Rubin is currently under full production, and the company has provided us with an extensive overview of the rack architecture, diving into individual components. When we talk about rack generations, NVIDIA is set to feature major upgrades with Vera Rubin, which we'll discuss in depth, but based on a recent video by CNBC diving into the Vera Rubin architecture, we saw an extensive look at multiple components, ranging from the main compute node to networking and cooling elements. More importantly, NVIDIA's Senior Director of Infrastructure, Dion Harris, calls Vera Rubin one of the "world's most complex AI systems, arguing that what NVIDIA does is unique and difficult to execute. Given that Rubin is expected to see customer commitments soon, it is important that we dive into what an NVL72 rack actually looks like. And, one of the most essential elements of the rack out there is, of course, the Vera Rubin SuperChip itself. We have already talked about how the Rubin GPU and Vera CPU configuration looks from a technical perspective, but one important point to note is that major performance improvements come from NVIDIA integrating HBM4 with the GPU, along with dedicated SOCAMM modules. Altogether, memory bandwidth reaches a whopping 1.2 TB/s. NVIDIA's major upgrade with Vera Rubin also comes within the cooling department, since Team Green plans to integrate modular liquid cooling designs, covering SuperChip elements such as Rubin GPU and Vera CPU, through dedicated cold plates. NVIDIA's executives argue that Rubin deployment will indeed convince hyperscalers to switch to upgraded liquid-cooling systems, and, interestingly, the current implementation reduces water use, another benefit touted by NVIDIA. NVLink is an important aspect of Vera Rubin NVL72, and with the 6th-generation interconnection fabric, often called the "NVLink Spine", NVIDIA plans to deliver a total aggregate bandwidth of 260 TB/s per rack. Harris says that with the latest NVLink generation, the company has taken modularity to a whole new level, which is why it claims the NVLink 6 spine supports zero-downtime maintenance and rack-level RAS services. While estimates suggest that Vera Rubin will debut with a decent price hike, NVIDIA says that the architecture brings in a 10x reduction in inference token cost and 4x reduction in the number of GPUs to train MoE models vs Blackwell GB200, which means that the "most you buy, the more you save" rule by NVIDIA's CEO is still intact.